Part IThe Story
A Gong Cast for One Listing
At 9:30 on the morning of July 9, 2018, Lei Jun struck a gong that had not existed a few months earlier. Exchanges keep instruments for this purpose the way churches keep bells, and for most companies the house gong in Hong Kong is good enough. Xiaomi was going to be the first company to list there under a "weighted voting rights" structure, the arrangement in which a founder's shares carry more votes than the public's. The Hong Kong Exchanges and Clearing commissioned a new copper gong for the occasion. It was cast for a company that was eight years old.
The ceremony was the clean part. Everything around it was less clean. The roadshow had opened on June 21 with Lei and his senior executives working the room in Hong Kong. The prospectus offered 2.18 billion shares at HK$17 to HK$22 and aimed to raise as much as US$6.1 billion, which implied a market value of $53.9 billion to $69.7 billion. Four years earlier, in December 2014, private investors had already priced the company at more than $45 billion. So at its very highest the public market was offering a modest premium on a number Xiaomi had reached while it was still selling phones only online. The day after the roadshow began, the Financial Times ran a headline saying Lei had hit "another wall in bid to take Xiaomi public." The offering priced at the bottom of the range and raised $4.72 billion. That is a fortune. It was also less than he had wanted.
Hold both of those in view: a new instrument cast in his honor, and a price set at the floor. Most of Lei Jun's career fits between them. He is perhaps the most admired founder in China and one of the most doubted. He spent years being called a "counterfeit Jobs" and then outsold Apple. The capital he has put at risk most often, and most publicly, is his own reputation. He has offered it as collateral for phones, for chips, and for a car, and the market has priced it at a premium and also, more than once, at the bottom of the range.
He was 48 when he struck the gong. He had spent sixteen years at a software company fighting Microsoft without winning, sold an online bookstore to Amazon for an amount that later looked small, spent several years as an angel investor watching other people's companies grow, and at 40 started a phone company in a market everyone agreed was already full. The new gong was the first instrument anyone had made specifically for him. He had made the first one himself, decades earlier, from batteries and a wooden box.
By the Numbers
The Lei Jun Ledger
$75MPrice Amazon paid for Joyo.com in 2004
$1M → $129MLei's 2005 stake in YY, valued at its 2012 IPO
40Lei's age when Xiaomi was founded on April 6, 2010
5%Publicly pledged ceiling on Xiaomi's overall hardware net margin
CN¥365.9BXiaomi revenue in 2024, up 35% year over year
410,000+Xiaomi electric vehicles delivered in 2025, the second full year of sales
76 secTime to build one SU7 at full capacity in Xiaomi's Beijing factory
Fire in a Borrowed Book
Lei came to computer science more or less by accident. He enrolled in Wuhan University's computer science department in 1987 largely because his friends were studying it. The library is where the accident turned into a direction. On a shelf was a Chinese translation of Fire in the Valley, Paul Freiberger and Michael Swaine's history of the personal computer. It was an old 1980s printing with faded type and a clumsy translation. It told the founding story of Apple: two Steves, an April Fool's Day incorporation, fifty computers built in thirty days to fill a first order.
— Lei Jun, recalling his first year at Wuhan UniversityThat was a book from the 1980s — crudely printed, roughly translated by today's standards — but I was absolutely electrified. I stayed awake for days and nights in university reading it, dreaming even then of building a world-class company.
The first thing to notice is the order of events. The ambition came from a book, and from a foreign book read in a poor translation. The model of greatness he took on was assembled from secondhand parts: an imported story, rendered imperfectly into Chinese, read by a teacher's son from Hubei whose father, by one widely repeated account, earned $7 a month.
The second thing is how quickly he acted on it. He finished the credits for a four-year degree in two years. He lived mostly on scholarships and on paid work for his professors' projects. Decades later he told CCTV that at one award ceremony, holding a sizable scholarship, he silently promised that if he ever could, he would repay the university "a hundred, a thousand — ten-thousand — fold." With a classmate, Wang Quanguo, he set up a small software workshop they called Yellow Rose Studio. It produced an encryption tool, BITLOK, and an antivirus program called Immunity 90, which sold for about a million yuan. That was a remarkable sum for a student in early-1990s China. He also co-wrote programming books on DOS and Windows. He had a hard drive full of working code before he had a diploma.
Then he tried to do what the book described. In his final year he and some technically capable friends, Li Ruxiong and Wang Quanguo among them, founded a company called Sanse, or Tricolor. There were four co-founders with equal shares. They had skills and some contacts in Wuhan, and no idea what the company was for. They assembled computers. They wrote software. They traded electronic components. At one point they did printing. When things were worst, the partners reportedly survived by playing cards with the cafeteria cook to win meal tickets.
Lei later admitted, without much embarrassment, that the founding team had barely discussed raising capital, opening a bank account, or finding revenue. His reasoning had been simple: Jobs and Gates had succeeded right out of school, so why not him? The company failed quickly. Years later, after he had become the person young Chinese founders quote, he said he did not encourage college students to start businesses. He has never pretended this contradicts his own history. He seems to see it as the lesson of that history.
He graduated in 1991 and was assigned to a research institute under the Beijing aerospace bureaucracy. Assignment was still how China's new graduates were placed in those years. At the end of that year he met Qiu Bojun, a figure at a young software company called Kingsoft. In early 1992 he went to Kingsoft's research department in Zhuhai as an intern. In July he joined for good.
Pangu, or the Price of Fighting a Giant
Kingsoft's word processor, WPS, was so common in the DOS era that it seemed to be installed on almost every computer in China. That was the problem. DOS was ending, and Microsoft was arriving in China with Windows and Office. Lei was put in charge of the Beijing development department on August 15, 1992, and became general manager of Beijing Kingsoft in 1994. The company's answer to Microsoft was a Windows office suite with a name taken from Chinese creation myth: Pangu, the giant who separated heaven from earth.
They spent three years on it. Kingsoft moved nearly all of its programmers onto the project. Lei later said the team had "bet on almost all of our families," working day and night. In April 1995 Pangu shipped. Lei remembered that the team had prepared a celebration banquet in advance.
Sales came in below one-tenth of expectations.
He has told this story many times, most fully in his third annual speech in August 2022, a three-hour talk about the low points of his life. In that telling, the failure didn't stay contained to one product. WPS revenue fell too. By the end of 1995 and into 1996, Kingsoft Beijing had a little over 100,000 yuan in the bank and could not cover the next month's salaries. Paydays were the worst days. People left. He was 26 and responsible for all of it.
What came afterward was a slow recovery, not a turnaround. In November 1996 the team released WPS 97. Kingsoft pushed into categories Microsoft cared less about: a translation dictionary called PowerWord, Kingsoft Video Master, and a role-playing game, The Legend of Sword and Fairy, that became famous in China in its own right. On August 12, 1998, Lenovo invested in and restructured Kingsoft. That year Lei became CEO at 28.
Asked about it at Wuhan University's graduation ceremony in June 2023, he credited the promotion to circumstance more than merit. "A 28-year-old Kingsoft CEO," he said. "Really not because I was so great. Purely an accident." The board hadn't found the right person and put him in charge in the meantime. Maybe so. It is also the kind of thing a man says when he has spent his life being measured against a story about a genius and has decided, at least publicly, not to claim to be one.
Kingsoft was a long, losing war of position against an opponent with more money, more reach, and no need to lose. One profile described his sixteen years there as a career spent chasing a giant "that never blinked." Along the way he found small exits. In 2000 he founded Joyo.com, an online retailer of books and media, built while Amazon was still deciding what it wanted to be outside the United States. On August 19, 2004, Amazon announced it would acquire Joyo for about $75 million, and Joyo became the base of Amazon China. It was a real return and a real success, and over the following years, as Chinese e-commerce grew past any forecast, it looked increasingly like a sale made too early.
In 2007 he took Kingsoft public on the Hong Kong Stock Exchange. On December 20, 2007, he resigned as president and CEO, citing "health reasons." He has described the years before that IPO as among the most painful of his life. By his own account he spent close to a year doing very little, and believed his best working years were behind him.
He had spent sixteen years trying to build a Chinese Microsoft. He had a fortune, an exit, and an IPO. He did not have the company from the book.
What the Pig Knows About Wind
The phrase most closely tied to Lei Jun in China is not about phones. It is about pigs: zhan zai fengkou shang, zhu ye neng fei qilai. Catch the right tailwind and even a pig can fly. It came out of the Kingsoft years, out of the experience of being early, being late, and watching good work fail because of timing. In China it became a meme, then a motto, and then a justification for every speculative startup of the 2010s. Lei has spent a good deal of energy since then explaining that people misread it.
His version, as summarized from Xiaomi's Entrepreneurial Thinking, has two qualifications. Before the wind comes, the pigs that end up flying are usually well prepared: the "young" founders who succeed have typically had years of business training and deep experience in their industries. And a pig that reaches the wind without enough strength falls fast. The wind is not luck. It is the reward for having built wings during the years when there was nothing to fly into.
He spent those years in an unusual place for a man with his drive: on the sidelines, writing checks. After leaving Kingsoft he became one of China's most active angel investors and backed more than twenty companies. In 2005 he put $1 million into YY, a real-time voice and video platform. When YY went public in 2012 those shares were worth $129 million. He became chairman of UCWeb, the mobile browser company, in 2008. He invested in Vancl, an online fashion retailer run by Chen Nian, a friend of almost three decades. Chen Nian appears later in this story, in tears, on a livestream.
From the board seat, Lei watched mobile internet startups get funded, grow, and collapse. He studied failure patterns: teams with fear of missing out, companies that, like his own Tricolor, could do anything and so did nothing well. He also kept a habit from his Kingsoft years that looked like a hobby and wasn't. He bought every new phone that came out, used it, took it apart, and put it back together, until he understood handsets as an engineer and complained about them as a customer. "I'm a mobile phone fanatic," he said at Peking University in 2014. "I think about them every day."
He was preparing without calling it preparation. When he decided to build a phone, he would be 40, older than nearly every founder he had funded. He would have run a public company, sold a startup to Amazon, studied twenty others from the inside, and taken apart more handsets than most phone executives had held. Jobs, in the book, started in a garage at 21. Lei's garage was a sixteen-year detour.
In 2011 he co-founded Shunwei Capital to keep investing in e-commerce, social networking, and mobile. On July 7, 2011, he went back to Kingsoft as chairman. He returned, in a supervisory role, to the company that had nearly broken him, the same summer his new company was about to announce its first phone.
Millet, Rifles, and a Rabbit in a Lei Feng Hat
On April 6, 2010, Lei and six partners founded a company named after a grain.
The founders are worth describing, because together they look like a deliberate collection of institutional experience, each borrowed from a larger company. Lin Bin, born in February 1968, studied radio electronics at Sun Yat-sen University, earned a master's at Drexel in 1992, worked as a network engineer at ADP, and then spent eleven years at Microsoft rising from software design engineer to engineering director. He left Microsoft for Google and became vice president of the Google China Institute of Engineering. He had spent his career inside the American giants Lei had spent his career fighting. In 2010 he and Lei were both 41, both married, and both fathers of daughters (Lin had three, Lei two), and they had spent months imagining a phone with top specifications and a low price. Zhou Guangping came from Motorola, where he was senior director of the Beijing R&D center. Hong Feng was a senior product manager at Google China. Huang Jiangji was a principal development manager. Li Wanqiang had run Kingsoft Dictionary, the PowerWord product that helped keep Kingsoft alive after Pangu. Liu De, born in 1973, chaired the industrial design department at the University of Science and Technology Beijing. He had trained at the Beijing Institute of Technology and had just completed a master's in industrial design at the Art Center College of Design in the United States. That degree was awarded in April 2010, the same month Xiaomi was founded.
The roster included Google, Motorola, Microsoft, Kingsoft, and an American design school. It amounted to a set of transplanted practices. The question was whether the result would be an imitation or an original.
The name pointed to both. Xiaomi means millet, the plain grain of north China. In 2011 Lei gave it more meanings than a startup name needs. Xiao, small, referred to a Buddhist idea that a single grain of rice can be as great as a mountain, which he presented as a commitment to starting with small things instead of perfection. Mi, he said, stood for "Mobile Internet" and also for "mission impossible," the obstacles of starting the company. He also said he thought the name was cute. In 2012 he told Reuters it was about revolution. Others noted a further echo he did not dispute: "millet and rifles," xiaomi jia buqiang, the Communist phrase for the underequipped army that beat better-armed enemies. When Xiaomi later named a mobile processor "Rifle," the echo got louder.
The company's mascot, Mitu, is a white rabbit wearing an ushanka, the fur hat known in China as a "Lei Feng hat" after the soldier the Communist Party made a model of selfless service. In the original design the hat had a red star and the rabbit wore a red Young Pioneer scarf. The star was later replaced with the company logo. The brand drew on Silicon Valley's founding myth and the People's Liberation Army's at the same time, and it is not clear Lei saw any conflict.
The first product was not a phone. On August 16, 2010, Xiaomi released MIUI, a customized version of Android. It began with almost nothing: phone, text messages, contacts, and a home screen. Four functions, chosen with deliberate restraint. The team shipped updates weekly, on an internet company's schedule rather than a handset maker's. Engineers talked directly with users on forums. Lei has said many times that the operating system started with its first hundred users. MIUI drew its first serious attention not in China but on XDA, the American developers' forum, where Android enthusiasts installed it on phones made by other companies. Xiaomi's international business began as a software download before the company had made a single device.
Lei later reduced the approach to a seven-character formula, zhuanzhu, jizhi, koubei, kuai: focus, extremity, word of mouth, speed. It reads like a management slogan. In practice it described how a company built from other companies' people moved faster than any of them.
That December the company raised a $41 million Series A. On August 16, 2011, exactly one year after MIUI, Lei introduced the Mi 1: a dual-core Qualcomm processor, Android underneath, MIUI on top, priced at 1,999 yuan when an iPhone cost 4,999 yuan or more. Engadget summed it up as "a $310 Android with 1.5GHz dual-core SoC and other surprises." Xiaomi took 300,000 preorders in its first 34 hours. In less than a year it sold more than three million. In 2012 it sold 7.19 million phones and reported about $2 billion in revenue, all without print or TV advertising. Forbes described the user meetups as having "the feel of religious revivals." Lei had more than four million microblog followers and spent hours on MiTalk, Xiaomi's chat platform, collecting complaints himself.
One of the oldest videos in the company's history, circulated again in 2025 for its contrast, shows Lei in 2011 on the street trying to get strangers to look at his phone. Most of them don't.
The Counterfeit Problem
Then there were the jeans. Lei appeared at early Xiaomi launches in jeans and a dark shirt and used the rhythms of Apple's keynotes, including, in time, a closing "one more thing." Chinese media called Xiaomi "the Apple of the East" and called Lei Lei-bu-si, a pun on Qiao-bu-si, the Chinese rendering of Jobs. In June 2013 the New York Times put it more harshly: "In China, an Empire Built by Aping Apple." Lei was a "professed Jobs acolyte" building "a knockoff of one of the gods of American ingenuity." In another widely circulated phrase, he was a counterfeit Jobs.
His public response is one of the most revealing things he has said. He denied copying the dress code. In 2012 he told Forbes the casual clothes at launches were a plug for Vancl, his friend's e-commerce fashion site. It may well have been both a tribute and an ad. On the comparison itself he said: "I was annoyed in the beginning, very annoyed. But I don't mind anymore." Associates said he enjoyed the nickname. Then, in the same interview, he did something no acolyte does. He explained why his model would have failed in his market.
— Lei Jun, interview with Forbes, 2012If Jobs had lived in China, I think he could not have succeeded. Jobs was a scrupulous perfectionist, while Chinese culture emphasizes the middle path.
In China, he added, "you also need to make compromises." This is the hinge of the whole story. The Jobs costume was real, and so was the admiration, but what Lei had actually built was close to Apple's opposite. Apple sold scarcity at a premium. Xiaomi sold abundance near cost. Jobs refused to compromise as a matter of identity. Lei treated compromise as a craft. Decades later, speaking to CCTV, he described the arc: "But the more I understood Jobs, the clearer it became that I'm not him."
The difference showed up most clearly in margin. Lei made the company's thin hardware margins a public promise, eventually stating that Xiaomi "will never raise the overall net profit margin on hardware above 5 percent." No mainstream hardware company says this. A margin is normally private, a number managed in the dark. Lei announced his like a vow. The bet was that phones priced near cost, sold directly online without middlemen, and kept on sale for about eighteen months (longer than most rivals, so component costs could fall underneath them) would win a user base that internet services could later monetize. The vow also served as a brand. It meant: we are not taking advantage of you. Lei flew economy and wore a watch that reportedly cost a little over 1,000 yuan. The personal frugality and the corporate margin promised the same thing.
The middle path had a cost, and the cost went on record. To manage demand and inventory, Xiaomi relied on flash sales: small batches, announced in advance, gone in seconds. Critics called it "hunger marketing" and said it fed on the cult of Apple. In 2014 Taiwan's Fair Trade Commission looked at the numbers. Xiaomi had advertised 10,000 units in each of its first two Taiwanese flash sales and 8,000 in the third. Actual sales were 9,339, 9,492, and 7,389. In the first sale the company had issued 1,750 priority "F-codes" that let some customers skip the public queue. The commission fined Xiaomi NT$600,000, about $20,000. It was a trivial sum and an awkward fact, because the flash sales had been sold as the purest form of the company's honesty, and the queue turned out to be partly staged.
Other complaints fit the same pattern. Android developers spent years criticizing Xiaomi for delaying the kernel source code it was required to publish under the GPL. The company promised the code in 2013, and a complete repository did not appear on GitHub until 2020. Reviewers kept praising products while noting what they resembled. Ars Technica's 2014 headline on the Mi 4 called it "China's iPhone killer," "unoriginal but amazing." Both adjectives were accurate. The company had become very good at making things that were not new.
— Lei Jun, to the New York Times, 2013We're making the mobile phone like the PC, and this is a totally new idea. We're doing things other companies haven't done before.
The Times treated that line as overreach. It is worth reading again. A PC was a commodity made of standard parts from many suppliers and differentiated by price and software. Lei was saying, accurately, that he was doing to phones what Dell and Lenovo had done to desktops. The claim was not that the phone was original. The claim was that the business model was. Most people heard the first.
Unicorpse
For a while the costume didn't matter, because the numbers were so large. Xiaomi sold 18.7 million phones in 2013 and more than 60 million in 2014, and became the largest smartphone seller in China. In the third quarter of 2014 it was third in the world for the first time. In August 2013 it hired Hugo Barra, Google's vice president of product management for Android, to run its international expansion. It was the first time a phone company had taken a senior executive from Google's Android team, and the press treated it as a statement of intent. In February 2014 Xiaomi opened an international headquarters in Singapore. In April it paid $3.6 million for the domain mi.com, the most expensive domain purchase in Chinese history. In December it raised $1.1 billion at a valuation above $45 billion, which made it one of the most valuable private technology companies in the world. In April 2015 Ratan Tata, the patriarch of Indian industry, bought a personal stake.
India showed both the opportunity and the risk. On December 9, 2014, the Delhi High Court granted Ericsson an ex parte injunction banning the import and sale of Xiaomi devices over a patent dispute. A week later the court allowed sales of Qualcomm-based devices to resume. On December 30, the Redmi Note 4G went on sale on Flipkart and sold out in six seconds.
Then the formula stopped working. The online-only model that had made Xiaomi suddenly strong also capped it. Rivals copied the software, improved the hardware, and built the physical retail networks that Chinese consumers in smaller cities still relied on. Phones still made up 94 percent of revenue in 2014. Xiaomi missed targets and supply chain failures piled up. Between 2014 and 2016 it fell from first to fifth in China and dropped out of the global top five. Analysts started using the word unicorpse: a billion-dollar startup presumed dead. No consumer hardware company, the argument went, recovers from a fall that steep.
Lei's response was the opposite of the delegation that founders usually describe in hindsight. On May 17, 2016, he personally took over the smartphone department. He reorganized the hardware team, which doubled in size, and strengthened research, supply chain, and quality management. He gave up the founding article of faith that Xiaomi would sell only online. "Mi Home" stores, 54 of them by the end of 2016, were turned into full retail outlets. Three of them, in Shanghai, Beijing, and Zhengzhou, each did more than 100 million yuan in sales that year. His January 2017 letter to employees set a target of 200 more stores in 2017 and 1,000 within three years. The company expanded its ecosystem of affiliated hardware makers and invested in companies like Roborock, in which it bought a 24.7 percent stake in 2014. A pre-crisis product, the Mi Band fitness tracker, had already made Xiaomi the world's second-largest wearables maker in early 2015, with 2.8 million units in a single quarter. By 2016, revenue from ecosystem products, TVs, and routers had passed 15 billion yuan, with more than 50 million connected devices.
The 2017 letter describes the year in the gentle language of a man still in the hole: "a year in which we took time to catch our breaths and slow our rapid pace to make essential adjustments." It also counted progress the critics had missed: more than 16,000 patent applications worldwide and 3,612 granted. The Mi MIX, released in late 2016, removed nearly every top-mounted component to reach a 91.3 percent screen-to-body ratio. It replaced the proximity sensor with ultrasound and the earpiece with a piezoelectric speaker that used the metal frame to produce sound. It was the first Xiaomi phone nobody could plausibly call a copy.
India carried the company through the recovery. On October 19, 2016, Lei sent an internal letter: Xiaomi had sold more than a million phones in India in eighteen days. He thanked "Manu and Hugo," meaning Manu Kumar Jain, who ran Xiaomi India, and Barra, and predicted Xiaomi would lead the Indian market within three to five years. It took about one. In the third quarter of 2017 Xiaomi passed Samsung to become India's largest smartphone brand, with 9.2 million units in the quarter. Barra had left by then, in early 2017.
The full recovery took five years. In the third quarter of 2020 Xiaomi was back to third in the world. In the second quarter of 2021, according to Canalys, it passed Apple with a 17 percent share and became number two globally for the first time. Lei's letter to staff began with something like disbelief: "For the first time ever, Xiaomi became No. 2 in the world." He then wrote about the fall plainly: "In 2014 Q3, we took third place globally for the first time in our history. Unfortunately, we ran into many difficulties subsequently and dropped out of the Top 5." The tenth anniversary of the first phone was a month away.
The List That Built a Car
In 2019 Lei received an honor that the Chinese state gives to private businesspeople it wants to signal approval of: he was named an "Outstanding Builder of Socialism with Chinese Characteristics." He had been a deputy to the National People's Congress since 2013, first for Guangdong and later for Beijing. In China this is a normal position for a founder of his prominence. It shows he is inside the system and expected to stay there.
On January 14, 2021, five days before the end of Donald Trump's first term, the U.S. Department of Defense listed Xiaomi as a company "owned or controlled" by the People's Liberation Army, which barred American investors from holding its shares. The Wall Street Journal reported that the evidence included the 2019 award. Two different countries were reading the same citation: in Beijing it meant the state approved of him, and in Washington it meant the state controlled him.
Xiaomi sued on February 17. On March 12 a federal judge blocked the ban, questioned the government's national security claims, and called the designation "deeply flawed." In May the Defense Department agreed to remove Xiaomi from the list. On paper Xiaomi won outright. In practice the episode had shown Lei how exposed a phone company is when it depends on American chips, an American-licensed operating system, and American capital markets.
On March 30, 2021, between the lawsuit and the settlement, Xiaomi announced it would build cars. Three years later, in July 2024, Lei said directly that the sanctions list was the reason. The announcement was a letter, and it was written as a personal commitment, not a corporate one.
— Lei Jun, open letter to Xiaomi employees, March 30, 2021I will once again lead the team by myself, serving as the CEO of the smart electric vehicle business. Just a while ago, at the launch event, I affirmed to our Mi Fans and everyone that it will be the last and important startup project in my life. I am willing to bet my lifelong reputation to fight for Xiaomi vehicles!
The rest of the letter is more modest than the headline. The car industry, he wrote, "features great complexity, a huge amount of investment, a long product cycle, and low fault tolerance." Xiaomi would approach it "with a fearful heart as if we are walking on thin ice." Rivals had "kick-started this business six years earlier than us." The commitment was $10 billion over ten years, starting with 10 billion yuan. One Chinese headline that week reduced his argument to two words: kui de qi, "we can afford the losses."
The day after the car announcement, Xiaomi unveiled a new logo by the Japanese designer Kenya Hara. The orange square became a "squircle" with rounded corners. The letters "MI" stayed identical. The orange got slightly darker. The work had reportedly been underway since 2017. Vice's headline: "Xiaomi Spent 3 Years To Create a New Logo That Looks Just Like the Old One." The internet laughed, and the joke was fair. Looking back, the pairing was close to self-parody: one day the company bet on the hardest manufacturing business in the world, the next it changed its corners. A company that had once made its name with a square had decided to soften it. Lei, meanwhile, was trying to build cars.
He started by learning to drive like an engineer. He later said he test-drove more than 170 cars during development and earned a racing license. In August 2021 Xiaomi bought an autonomous-driving company, Deepmotion, for $77 million. In February 2022, with Xiaomi's share of the Chinese phone market slipping from 17 percent to 14 percent to 13.2 percent over three quarters, he posted on Weibo that the company would benchmark itself directly against Apple in premium phones, spend $15.7 billion on R&D over five years, and treat the effort as a "life-or-death battle for our development." That year Leica, the German camera maker, moved its smartphone partnership from Huawei to Xiaomi. A year after an American blacklist, a German heritage brand made the opposite judgment.
On December 28, 2023, Xiaomi unveiled the SU7, a sleek electric sedan, and announced a goal of becoming one of the world's five largest automakers. It went on sale in Beijing on March 28, 2024, starting at 215,900 yuan. At first Xiaomi lacked its own production license, so the car was built under contract with BAIC Group. It received the license in July 2024. The factory in the Beijing Economic-Technological Development Area is built around a proprietary die-casting system with the plain name Hyper Die-Casting 79100 Cluster, and at full capacity it produces an SU7 every 76 seconds.
In May 2025 Xiaomi released the XRING O1, a self-designed 3-nanometer mobile processor, which made it, by Chinese accounts, the fourth company in the world able to design one. The chip program dated back to the Surge S1 project of 2014, an effort widely written off in its early years. The pressures of a decade had pushed a company built on assembly into building cars and chips.
The Wishing Well and the Bill
Lei sold the car the way he had sold the first phone, by putting himself in front of it, now with far more reach. In 2024 he went live on Douyin to talk about the SU7's first twenty days on sale, and more than 100,000 viewers joined within a minute. After two hours, the stream had more than 180 million likes. A May livestream for the SU7 Pro drew 39 million viewers by his count. On December 31, 2024, he hosted a New Year's Eve stream from the car factory that logged 37.884 million views and peaked at 496,000 concurrent viewers. That month, he announced, SU7 deliveries had passed the Tesla Model 3. In 2024 Xiaomi reported 365.9 billion yuan in revenue, up 35 percent. Its new businesses, mostly cars, brought in 32.8 billion yuan. The share price quadrupled in a year. In February 2025 Lei was briefly the richest person in China, and he was among the entrepreneurs at President Xi Jinping's February symposium with private business leaders.
His standing had become something like moral authority. In March 2025, after China's annual consumer-rights broadcast exposed sanitary pads that failed hygiene standards, thousands of users went to Lei's social media accounts and asked Xiaomi to make sanitary pads, a business it has never been in. He described his position with some bemusement.
"Whenever something seems difficult to achieve, or when any product or service loses people's trust, people come to my comment section to leave messages," he said. "I've become a kind of wishing well."
A wishing well holds coins people throw in. It does not owe any of them anything. What Lei had built was closer to a bank, and banks get runs.
The run started on a highway in Anhui province in late March 2025. An SU7 with Xiaomi's driver-assistance system engaged was traveling at 116 kilometers an hour when the system alerted the driver to take over, about two seconds before the car hit a barrier. Three people died. The public questions were about the technology, which in China is legally Level 2 and requires the driver to stay in control, and also about the marketing, which critics said encouraged drivers to forget that. In April the Ministry of Industry and Information Technology summoned representatives from sixty companies and warned them not to overstate driver-assistance capabilities. Lei, who had posted constantly in late March, posted 21 times in April, almost entirely official company updates. In May he wrote that the past month had been the most difficult since he founded Xiaomi in 2010. He had canceled meetings and trips and stepped away from social media. The post became Weibo's top trending topic.
More problems followed. Owners of the SU7 Ultra organized online over a carbon-fiber hood whose ventilation openings, they said, didn't do what the marketing suggested. In September 2025 Xiaomi recalled 117,000 SU7s to update the high-speed assisted-driving system. In November a court in Haikou ruled on a dispute over an Aurora Purple SU7 Max: Xiaomi had unilaterally scheduled the car for production and demanded the buyer pay the remaining 313,900 yuan within seven days, before she could inspect it, and then kept her 5,000-yuan deposit when she couldn't pay. The court called the relevant contract terms "unfair and unreasonable" and ordered Xiaomi to pay back double the deposit: 10,000 yuan. In the separate hood case, Xiaomi's lawyers argued that Lei himself had repeatedly said the hood option was "not recommended to buy," and that his posts about the hood's cooling had been deleted, so the company had not intentionally misled anyone. Chinese media described the legal department as having stabbed its chairman in the back. Over one month that autumn Lei lost about 699,000 followers. Xiaomi's market value fell by more than HK$550 billion from its high.
None of this stopped the sales, and that was the strange part of 2025. When the YU7, Xiaomi's first SUV, launched on June 26, it took more than 200,000 preorders in three minutes and more than 240,000 locked orders within eighteen hours. Xiaomi delivered more than 410,000 cars that year. The trust being drawn down was still very large.
On September 25, 2025, Lei gave his sixth annual speech, titled Change. Beforehand he posted a photo of himself in a T-shirt that said "It's the age to take risks." In the speech he explained that the slogan came from Chen Nian, the Vancl founder and his friend of 28 years, who was starting over in live-streamed e-commerce. Chen Nian was broadcasting on Douyin at that moment. When Lei mentioned him, Chen covered his face and cried. In an interview the same night Lei offered a philosophy of management that sounded more like a warning to himself: "There are often intense and heated debates within Xiaomi, even over a single issue for a whole year. But this way, the boss is less likely to make mistakes."
On January 7, 2026, he went live again and took a car apart on camera. He showed off the 2025 delivery numbers. He demonstrated a triple-redundant power design to answer rumors that the car's doors wouldn't open after a crash. He promised to end the industry's use of fine-print disclaimers. He addressed the "marketing master" label he had carried for a decade, explaining that it came from a 2013 or 2014 TV competition in which a rival team led by Liu Qiangdong of JD.com joked that nobody should try to out-market Lei Jun, and that competitors had since turned the joke against him. He said, smiling, that he would like to talk it over with Liu sometime. He disciplined the public relations department for how it had handled a dispute with an online critic, and asked fans to give it a chance to correct its mistakes. Xu Jieyun, the head of public relations, answered a question that had followed the company for years: Lei once said Xiaomi didn't advertise, and now it did. When the company was young, she said, engineers talked to users directly and there was no marketing department. Today it has 70,000 to 80,000 employees.
The financial results came in the summer of 2026. First-quarter revenue was 99.142 billion yuan, down 10.9 percent and below 100 billion for the first time in six quarters. Net profit attributable to shareholders fell 56.5 percent. Smartphone shipments dropped 19.2 percent to 33.8 million units, and phone gross margin fell to 10.1 percent. In mainland China, Xiaomi fell out of the top five, while Huawei, with its own chips and operating system, returned to the top of the market. The division that includes cars lost 3.1 billion yuan in operating profit. The shares, which had peaked at HK$61.45 on June 27, 2025, closed at HK$21.30 on June 26, 2026, down 65 percent in a year, while Chinese tech stocks overall fell about 16 percent. Analysts said what they had said in 2016: the formula had reached its limits, and the company relied too heavily on its founder.
It was a second unicorpse, this time much larger. Lei was 56. He had said the car would be his last startup.
One Yuan Short
In 1997, while Kingsoft was still recovering from Pangu, Lei gave 140,000 yuan to Wuhan University. It was the first installment on the promise he had made silently at a scholarship ceremony. In 2016 he gave the university money for a science and technology building. The certificate read 99,999,999 yuan, one yuan less than 100 million. Asked why, he said he wanted his gift to come in just under that of a fellow alumnus, Chen Dongsheng, whom he called his elder brother, and that he hoped more alumni would give instead of competing over amounts. This was the middle path written into a dollar figure: a donation large enough to matter and calibrated to avoid embarrassing anyone. A perfectionist would have rounded up.
In 2021, the year China's leadership began promoting "common prosperity" and its tycoons began giving accordingly, Lei donated more than $2.2 billion worth of Xiaomi shares to charity. In November 2023, for the university's 130th anniversary, he gave Wuhan University 1.3 billion yuan, about $182 million. It was the largest cash gift a Chinese university had ever received from an individual alumnus. He did not round anything down this time.
That June he had given the graduation address at Wuhan University, to more than 17,000 students, the largest class in the university's history, including graduates from 2020 whose ceremonies had been canceled during the pandemic. He told them about the library and Fire in the Valley. He told them that becoming CEO at 28 had been an accident. And he told them about the first lesson he received at the university, from an elderly professor speaking to freshmen. Some students had been placed in the chemistry department, the professor said, and were worried they would never find work. They were wrong to worry. What matters most at a university, he said, is not the knowledge you learn but learning how to learn. If you can learn, you can do anything. Lei said that sentence carried him through the rest of his life.
Xiaomi's president today is Lu Weibing. He was born in 1975 and earned a bachelor's degree in chemistry from Tsinghua University in 1998. He spent nine years selling phones at Konka, ran overseas business at Tianyu, served seven years as president of Gionee, a phone maker that faded, and founded his own company, Chenyee, which lasted about a year. He joined Xiaomi in 2019, when he was 43, as one of the industry's well-traveled survivors. He now runs the smartphone business. His degree is in chemistry.
In 2024 Wuhan University started what it calls the Lei Jun Class: thirty computer science students a year, half admitted through the national entrance examination and half chosen from first-year students, with flexible credit requirements, rotations through research labs, and a guaranteed path to graduate school for those who qualify. When the first cohort visited Xiaomi's technology park in late 2024, Lei told them, "In my heart, Wuhan University is the best university in the world." He told them computer science was still a field where a person could change the world. He told them to absorb as much as they could, like sponges.
If one of them goes home to the place where he was born, Group 4, Zhaowan Village, Louhe, in Xiantao, Hubei, on the flat farmland west of Wuhan, they will find a place where, according to the story his biographers have repeated so often it has become part of the legend, the first electric lamp was not purchased. A schoolteacher's son built it, with his father's encouragement, from parts made by other people: two batteries, a bulb, some wires, and a wooden box he made himself.
For the full early history, see Lei Jun's Cards: Secrets Behind the Miraculous Rise of Xiaomi; Chinese readers can consult the Biography of Lei Jun.
Part IIThe Playbook
The principles below come from the record in Part I: the decisions Lei Jun made, the ones he reversed, and the ones that cost him. Several of them conflict with each other, which is accurate. His career shows a man adjusting between imitation and originality, between thin margins and expensive bets, and between the leverage of a public persona and the damage it can do. These are not rules to copy. They are patterns, each with its conditions and its price.
Twelve Principles from Lei Jun
- 1.Borrow the script, then rewrite the lines.
- 2.Build the wings before the wind arrives.
- 3.Ship four functions, not forty.
- 4.Publish your margin as a promise.
- 5.Assemble founders from other people's institutions.
- 6.When the formula breaks, take the wheel yourself.
- 7.Convert an external shock into a mandate.
- 8.Enter late with fear, not swagger.
- 9.The founder is a channel. Channels carry liabilities.
- 10.Argue for a year so the boss is wrong less often.
- 11.Know which compromises are the middle path and which are fine print.
- 12.Repay the institution that formed you, at scale.
Principle 1
Borrow the script, then rewrite the lines
Lei's ambition came from an imported book, and his early public persona came from Apple's keynote style. He never denied the debt. He kept the jeans, the stage, and eventually "one more thing." What he changed was the economics. Apple sold scarcity at a premium, and Xiaomi sold abundance near cost. The New York Times saw the costume and called him a knockoff. It missed the business model, which was the opposite of the one being imitated.
The lesson is not "copy the leader." Imitation works as an apprenticeship when you know exactly which parts you are borrowing and which you are rejecting. Lei could say why Jobs would have failed in China (the perfectionism, the refusal to compromise), and that is evidence he had studied the model closely enough to see its limits. The borrowed form gave Xiaomi immediate recognition. The rejected substance gave it a strategy.
There is a cost. The "counterfeit" label stayed with Xiaomi for a decade and shaped how Western reviewers and investors saw it well after the products had become original. Borrowed charisma comes with the original's critics attached.
Tactic: Write down the three things you are deliberately copying from your role model and the three you are deliberately rejecting, and make the rejected three the core of your strategy.
Principle 2
Build the wings before the wind arrives
The "pig in the wind" line is usually quoted as an argument for timing. Lei's own reading is about preparation: the pigs that fly were ready before the wind came. His record supports this. He founded Xiaomi at 40, after sixteen years running a public software company, an exit to Amazon, and years on the boards of twenty-plus startups. He had also spent years taking apart every new phone on the market. When mobile internet arrived, he already understood handsets as an engineer, a customer, an investor, and an operator.
His first company, Tricolor, is the counterexample. It had talent and no preparation: four equal partners, no focus, no bank account plan, partners winning meal tickets at cards. His later advice that college students should generally not start companies is not hypocrisy. It is what he concluded from that failure.
L
Five Reinventions, One Career
Each phase supplied a capability the next one required.
1992
Joins Kingsoft. Learns to ship software against Microsoft and survives Pangu's failure.
2000
Founds Joyo.com. Learns e-commerce and direct-to-consumer logistics; sold to Amazon in 2004.
2007
Leaves Kingsoft as CEO. Begins angel investing in YY, UCWeb, Vancl and others; studies how startups fail.
2010
Founds Xiaomi at 40. Combines software iteration, online retail, and handset knowledge.
2016
Takes over the smartphone unit personally. Adds offline retail and an ecosystem strategy.
2021
Announces EV business after U.S. blacklisting; becomes CEO of the car unit himself.
2025
Releases the XRING O1 3nm chip; delivers 410,000+ cars.
Tactic: Before waiting for a market to open, list the specific capabilities you would need on the day it does, and spend the waiting period acquiring them, ideally on someone else's payroll or balance sheet.
Principle 3
Ship four functions, not forty
Xiaomi's first product was not a phone. It was MIUI, launched with four functions: calls, texts, contacts, and a home screen. It was updated weekly and shaped by conversations with its first hundred users on forums. That restraint is the first word of Lei's seven-character formula (focus, extremity, word of mouth, speed), and the other three depend on it. You can't iterate weekly on forty functions. You can't build devoted word of mouth with a product that is mediocre at everything.
The approach also produced Xiaomi's first international audience almost by accident. MIUI found early users on XDA, an American developer forum, and they installed it on phones Xiaomi didn't make. A small, excellent product spreads by itself. A large, average one needs advertising, which Xiaomi famously did not do at first.
Tactic: Cut your launch scope to the smallest set of features your first hundred users touch daily, and commit to a public update schedule shorter than any competitor's.
Principle 4
Publish your margin as a promise
Most companies treat margin as a secret. Lei made it a vow: Xiaomi would never raise its overall hardware net profit margin above 5 percent. The promise did several jobs. It positioned the company against premium rivals. It gave customers a reason to trust a new brand. It made the price itself the marketing. It also matched his personal reputation for frugality (economy flights, a 1,000-yuan watch), so the corporate promise and the founder's life told the same story.
The cost appeared in 2026, when analysts blamed exactly this low-margin, price-led model for leaving Xiaomi with no cushion against rising component costs and a Huawei that owned its chips and operating system. A public margin cap is a strong trust signal and also a strategic constraint you have announced to your competitors. The move works when the services layer above the hardware really does monetize, and becomes a trap when it doesn't keep up.
Tactic: If you publish a pricing or margin commitment, publish the business logic that makes it sustainable as well, and review annually whether that logic still holds.
Principle 5
Assemble founders from other people's institutions
Xiaomi's seven founders brought experience from Google, Microsoft, Motorola, Kingsoft, and an American design school. Lin Bin brought eleven years of Microsoft engineering management and Google's China institute. Liu De finished his Art Center design degree the month the company was founded. Li Wanqiang had run one of Kingsoft's few profitable products. None of them had built a phone company, and together they had seen most of how one works.
The same habit continued with hires. Hugo Barra came from Google's Android leadership to run international expansion. Lu Weibing, who now runs the smartphone business, arrived in 2019 after running Gionee, a rival that had declined. Lei repeatedly hires people whose earlier institutions taught them something specific, including people whose companies failed, which often teaches more.
M
Conventional Wisdom vs. the Lei Jun Approach
Where Xiaomi's practice diverged from standard startup advice.
| Conventional wisdom | Lei's approach | Outcome |
|---|---|---|
| Start young, start lean. | Start at 40 with veterans from Google, Microsoft, Motorola. | Scale |
| Keep margins private. | Publicly cap hardware net margin at 5%. | Trade-off |
| Launch a full product. | Launch an OS with four functions, update weekly. | Scale |
| Founders delegate operations as they scale. | Personally take over the failing phone unit in 2016. | Recovery |
| Don't make the founder the brand. | Livestreams with tens of millions of viewers. | Risk |
| Enter new categories with partners. | Build cars in-house after a contract-manufacturing bridge. | Unresolved |
Tactic: When recruiting co-founders, map each candidate to the specific institutional knowledge they carry, and decline anyone whose main asset duplicates your own.
Principle 6
When the formula breaks, take the wheel yourself
Between 2014 and 2016 Xiaomi fell from first to fifth in China and was being called a unicorpse. Lei did not hire a turnaround executive. On May 17, 2016, he took over the smartphone department himself. He then dropped the founding belief that Xiaomi would sell only online, opened physical stores, doubled the hardware team, and expanded the ecosystem.
What matters is that he gave up the formula and kept the team. The online-only model had been part of Xiaomi's identity, and dropping it meant admitting the original insight had an expiration date. The recovery took five years, from collapse in 2016 to number two worldwide in 2021, and the 2021 letter describes the drop without spin. A founder who can write "we ran into many difficulties subsequently and dropped out of the Top 5" in a celebration memo has kept an honest internal record.
Tactic: When results collapse, separate the formula from the people, and if the formula is what failed, take personal operating responsibility for replacing it instead of delegating the replacement.
Principle 7
Convert an external shock into a mandate
The U.S. blacklisting of January 2021 lasted four months, and Xiaomi won in court. Lei still treated it as a strategic warning, not a legal problem. By his own account in 2024, it was the reason Xiaomi built cars. The episode showed that a company dependent on foreign chips, foreign software, and foreign capital could be disrupted by a single government order. The response was vertical expansion: cars, then a 3nm chip, then foundation AI models.
The skill is to read a shock for what it reveals, not just for what it costs. A lawsuit fixed the immediate problem. The car program addressed the vulnerability underneath it. Whether the response was proportionate is still unclear, given the auto division's 3.1-billion-yuan quarterly loss in early 2026. It was, however, a response to the actual threat rather than the visible one.
Tactic: After any external shock you survive, write one sentence naming the dependency it exposed, and fund at least one initiative that reduces that dependency even if the immediate crisis is resolved.
Principle 8
Enter late with fear, not swagger
Lei's car announcement is often remembered for its boldness: ten billion dollars, a "last startup," a lifelong reputation at stake. The letter itself is mostly about fear. He acknowledged that rivals were six years ahead and that the industry had "low fault tolerance," and he described the company as walking on thin ice. He then laid out six specific advantages Xiaomi brought to the market, including cash reserves large enough to "invest boldly without hesitation."
The execution matched that tone, at least at first. He test-drove more than 170 cars. Xiaomi bought an autonomous-driving company. It built its first cars under contract with BAIC before it had its own license. It invested in proprietary die-casting. The late entrant studied before building. Where trouble came later, in assisted driving and the hood dispute, it came from marketing getting ahead of engineering, the swagger the letter had warned against.
Tactic: When entering a market late, write the incumbents' head start in years at the top of your plan, and require every marketing claim to be checked against it.
Principle 9
The founder is a channel. Channels carry liabilities.
Lei's personal following (tens of millions on Douyin and Weibo, livestreams with nearly 40 million viewers) let Xiaomi skip the hardest phase of building trust in a new category. Many SU7 buyers bought because they trusted him. That is a very efficient distribution channel, and Xiaomi saved enormously on advertising by using it.
A channel built on one person carries that person's risks. After the 2025 crash, the hood dispute, and the Haikou ruling, Lei lost nearly 700,000 followers in a month, and analysts identified dependence on the founder as a structural weakness. The low point came when Xiaomi's own lawyers cited Lei's statements and deleted posts as a legal defense, a sign that the company and its founder's voice had become partly separate. His "wishing well" comment states the problem: people bring their hopes to the founder, and when the product falls short, the founder carries the disappointment as well.
Tactic: If your founder is a primary marketing channel, build a clear review process for every public claim the founder makes, because their posts will be treated as product specifications.
Principle 10
Argue for a year so the boss is wrong less often
In September 2025 Lei described Xiaomi's internal culture as one of long, heated debate, sometimes over one issue for a whole year, and defended it by saying, "this way, the boss is less likely to make mistakes." The car decision followed "numerous rounds of debates among the executive team," according to his 2021 letter.
This complements Principle 9. A founder with tens of millions of followers will find that few people inside the company want to contradict him. Building deliberate, slow internal disagreement is how a charismatic leader protects himself from his own following. The cost is speed, and Xiaomi was built on speed. Lei appears to have concluded that for decisions involving physical safety and ten-year capital commitments, speed is the wrong thing to optimize.
Tactic: Classify decisions by reversibility, and require a formal dissent period, with named opponents, for every decision that cannot be undone within a year.
Principle 11
Know which compromises are the middle path and which are fine print
Lei told Forbes in 2012 that China requires compromise and that Jobs's perfectionism would have failed there. Much of Xiaomi's success comes from good compromises: near-cost pricing, long product cycles, borrowing Android instead of building from scratch, working inside the political system. The 99,999,999-yuan donation is the same instinct expressed in philanthropy.
Some compromises damaged trust. The Taiwan flash-sale counts, the 1,750 F-codes, the delayed GPL kernel releases, the Haikou deposit terms, and the fine print that Lei publicly promised in January 2026 to stop using all belong in the second category. The difference is who absorbs the cost. A middle-path compromise shares the cost between company and customer. A fine-print compromise moves it quietly to the customer. Lei's public promise to end small-print disclaimers reads as his recognition, late, that the two had blurred.
Tactic: For every compromise in your product or contracts, ask whether the customer would agree to it if it were stated in large print, and remove the ones they wouldn't.
Principle 12
Repay the institution that formed you, at scale
Lei's giving to Wuhan University runs through his whole career: 140,000 yuan in 1997, when Kingsoft was still recovering; 99,999,999 yuan in 2016; 1.3 billion yuan in 2023, the largest individual alumni gift in Chinese higher education; and in 2024, the Lei Jun Class, which selects thirty computer science students a year. He traces the commitment to a promise he made silently at a scholarship ceremony.
This is more than sentiment. It reinforces the founder's identity and builds a talent pipeline. The Lei Jun Class combines his name, his field, and his idea of how industry and academia should work together, and it produces students who will know Xiaomi before they know its rivals. The giving also aligns with the Chinese state's push for "common prosperity," which a founder who sits in the National People's Congress cannot ignore. The gift serves sincere motives and strategic ones at once.
Tactic: Identify the one institution that most shaped your ability to learn, and set up a recurring, named commitment to it that grows as your company grows.
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
— Lei Jun, on the Steve Jobs comparison, Forbes, 2012I was annoyed in the beginning, very annoyed. But I don't mind anymore.
— Lei Jun, CCTV interview, December 2023But the more I understood Jobs, the clearer it became that I'm not him.
— Lei Jun, Peking University, 2014I'm a mobile phone fanatic; I think about them every day. Building mobile phones is not my job, it's my passion.
— Lei Jun, on public expectations, March 2025Whenever something seems difficult to achieve, or when any product or service loses people's trust, people come to my comment section to leave messages. I've become a kind of wishing well.
— Lei Jun, interview with China Entrepreneur, September 25, 2025There are often intense and heated debates within Xiaomi, even over a single issue for a whole year. But this way, the boss is less likely to make mistakes.
Maxims
- The pig was ready before the wind. Timing pays off only for those who prepared during the years when nothing was happening.
- Copy the costume, reject the economics. Imitation is useful when you can say exactly where the model you're copying would fail.
- Four functions beat forty. A small product loved by a hundred users spreads faster than a large one tolerated by thousands.
- A published margin is a contract. Announcing what you won't earn builds trust and also tells competitors your limits.
- Hire the institutions, not just the people. Every co-founder should bring a playbook learned somewhere you have never worked.
- When the formula fails, the founder operates. A collapse calls for personal command, not a new hire.
- Read the shock for the dependency. Winning the lawsuit fixes the crisis; fixing the vulnerability is the real work.
- Enter late with fear in the memo. Write down the incumbents' head start and check every claim against it.
- A famous founder's posts are product specs. The bigger the following, the more each sentence is treated as a promise.
- The middle path is shared cost; fine print is hidden cost. Keep the first kind of compromise and get rid of the second.
