He Xiaopeng

The man who had already won #

He Xiaopeng (何小鹏) sold his first company for $4.3 billion, achieved financial freedom, and discovered something terrifying: emptiness. So he bet over $300 million of his own money on building cars — an industry he knew nothing about. His wife opposed it. His friends questioned it. Lei Jun (雷军), founder of Xiaomi and one of China’s most respected tech entrepreneurs, would eventually tell him that his biggest problem was himself.

A chairman can be a nice guy, but a CEO cannot.

The story of XPeng Motors is a story of technology, capital, and China’s most brutal industrial competition. But the story of He Xiaopeng is something rarer and more difficult to document: the story of a man who built an organization in his own image, watched that image nearly destroy the organization, and then dismantled both the image and the organization to build something that could survive without his personality holding it together.

The software executive’s gilded cage #

In 2004, Xiaopeng borrowed RMB 800,000 from Ding Lei (丁磊), the CEO of NetEase, and co-founded UCWeb with Liang Jie (梁捷). The mobile browser was a bet on a future most people in China had not yet imagined — that hundreds of millions of people would access the internet from their phones rather than their desktops. The bet paid off spectacularly. UCWeb grew to 500 million global users and attracted the attention of Alibaba, which acquired the company in 2014 for $4.3 billion in what was then the largest internet merger-and-acquisition deal in Chinese history.

Xiaopeng had won. He was financially free. He held a vice-presidential title at one of the world’s most powerful technology companies. And he was miserable. The Alibaba role was what he later described as a gilded cage: prestigious, comfortable, meaningless. Financial freedom, he discovered, felt “even more empty, even more pained” than the years of uncertainty that had preceded it. The man who had spent a decade building something from nothing found that having everything felt worse than having nothing at all.

Within months of the Alibaba deal closing, Xiaopeng was angel-investing in a small automotive startup in Guangzhou called Orange Mobility (橙行智动) — the company that would become XPeng Motors. By August 2017, he had made his decision. He resigned from Alibaba on August 22 and joined XPeng as chairman seven days later. He told Alibaba’s CEO, Zhang Yong (张勇): “All in for ten years.” His wife opposed the decision. Xiaopeng invested over $300 million of his own money into a company building cars — a discipline governed by physics, supply chains, and factory floors, none of which he understood.

“If I don’t put my entire fortune on the line,” he told his wife, “how will anyone trust me?”

The question was not rhetorical. It was the founding logic of his second career: personal financial exposure as a substitute for industrial experience. He was asking investors and employees to follow a software entrepreneur into manufacturing. The only way to make that credible was to make the stakes irreversible.

The dreamer’s organization #

The company Xiaopeng built from 2017 to 2022 reflected the man who built it. He was, by his own description and by the accounts of colleagues who spoke to Chinese journalists, a “lighthearted internet dreamer” (互联网浪漫主义者) — a leader who prized harmony, avoided confrontation, and believed that talented people would figure things out if you gave them enough freedom. In meetings, colleagues described his style with three words: uninformed, disagreeing, non-committal. He would question a proposal, express vague discomfort, and then let the proposal proceed because he did not want to overrule the people he had hired.

This leadership style had worked at UCWeb, where the product was software and the cost of a wrong decision was a bad release that could be patched in a week. Manufacturing is not software. A wrong decision about procurement, tooling, or vehicle configuration cannot be patched. It ships.

On September 21, 2022, XPeng launched the G9 — a flagship SUV that Xiaopeng had publicly called “the best SUV under 500,000 RMB.” Within hours, customers revolted. The trim configurations were incomprehensible. The base model at RMB 309,900 lacked adaptive cruise control. Key features — LiDAR-based driver assistance, fast charging — were locked behind expensive packages that made the vehicle’s true cost far higher than its advertised price. The very next day, competitor Li Auto (理想汽车) announced a rival SUV targeting the exact same price range. Chinese netizens observed that “XPeng was still dazed from falling flat, then Li Auto picked them up and gave them an over-the-shoulder throw.”

The G9 delivered 184 units in its first full month. The stock fell eighty percent that year. But the G9 was not the crisis. The G9 was the symptom. The crisis was Xiaopeng himself.

The call that changed everything #

Xiaopeng knew it. In the weeks after the G9 disaster, he called Lei Jun — Xiaomi’s founder, an early XPeng investor, and the closest thing Xiaopeng had to a mentor in the technology world. Lei Jun’s response was devastating in its precision: “XPeng’s biggest problem is that you only want to be chairman. You don’t want to be CEO.”

A chairman sets direction. A CEO makes people do things they do not want to do. Xiaopeng had spent his career being a chairman — even when his title said CEO. He made decisions by consensus. He avoided firing people. He gave talented executives freedom and hoped they would use it wisely. The G9 had proved, in the most public and expensive way possible, that hope is not a management strategy.

“I spent two years relearning how to be a CEO,” Xiaopeng told LatePost (晚点) in February 2025. The sentence is precise. Not “learning to be a better CEO.” Relearning. He was admitting that the identity he had built over two decades — the internet romantic, the consensus-seeker, the leader who wanted to be liked — was the root cause of the worst product failure in XPeng’s history. The G9’s baffling configuration matrix, the lack of quality control, the organizational dysfunction that no one had flagged — all of it traced back to a culture where no one told the boss he was wrong because the boss had designed the culture to avoid exactly that kind of confrontation.

The eighteen months from September 2022 to early 2024 were what Xiaopeng later called his “hell mode.” The stock collapse was painful. The market-share loss to Li Auto — which became profitable in 2023 while XPeng posted its worst quarterly loss since going public — was humiliating. Tesla’s January 2023 price war crushed XPeng’s sales by sixty percent year over year. Deutsche Bank warned that management was “making its last stand” with the upcoming G6 launch. Suppliers grew less accommodative. Rumours circulated that XPeng would not survive the year.

But the deepest wound was personal. Xiaopeng was not watching his company die from external forces. He was watching it die from the organization he had created — an organization that reflected his personality, his aversion to conflict, his belief that smart people would self-organize. They had not.

Swimming in blood water #

The man who emerged from the G9 crisis bore little resemblance to the one who had entered it. The dreamer became what colleagues called a “workshop foreman.” Xiaopeng personally took over procurement management. He worked six and a half days a week, arriving before ten in the morning and leaving after ten at night. A colleague told a Chinese journalist: “He is playing with his life.”

The organizational purge was the most aggressive in recent Chinese automotive history. Ten of twelve senior executives departed by September 2023. Eighty-five percent of top department heads were replaced. Ninety percent of frontline managers were rotated. In January 2023, Xiaopeng hired Wang Fengying (王凤英), the legendary “Iron Lady” who had spent thirty-one years at Great Wall Motors — an executive known for a management style so demanding that it was the opposite of everything Xiaopeng had previously valued. He was no longer hiring people he liked. He was hiring people who would challenge him.

In March 2024, he moved his own co-founders — Xia Heng (夏珩) and He Tao (何涛), the men who had started the company before Xiaopeng joined — to “lifetime honorary” positions. Industry observers noted that XPeng had “officially entered the He Xiaopeng era.” The observation was more precise than intended. The company had always been Xiaopeng’s. What changed was that he was no longer hiding behind consensus to avoid the loneliness of command.

The emotional cost was immense. In the second quarter of 2024, XPeng was still selling only a few thousand cars a month. Xiaopeng ate more than thirty dinners with over six hundred employees, personally making the case that the situation would improve by the fourth quarter. He described the experience to LatePost with a phrase that resists elegant translation: “吃得我吐血” — eating until he vomited blood, a Chinese expression for emotional and physical exhaustion pushed to the breaking point. About thirty percent of those employees left anyway.

“The hardest thing,” Xiaopeng said, “is making a team that has fallen into a valley regain confidence and rebuild their thinking. You do many things, but whether you can ultimately rebuild confidence — you can’t control that.”

He could not control it. But he could refuse to stop trying. The Volkswagen Group’s $700 million investment in July 2023 — made at the moment of lowest internal confidence — validated XPeng’s technology at precisely the moment Xiaopeng needed external proof that his bet was not delusional. The G6 launched at prices twenty percent below the Tesla Model Y, and the stock surged. By late 2024, the turnaround had materialized. Nearly forty thousand vehicles a month. In January 2025, XPeng became China’s top new-energy startup brand by monthly volume.

The ten years are nearly up #

In August 2017, Xiaopeng told Zhang Yong he was “all in for ten years.” The decade is nearly complete. From borrowed capital and a mobile browser to a declared artificial intelligence platform operating in forty-six countries, the trajectory is extraordinary. But the trajectory is not the story. The story is what Xiaopeng had to destroy to get here.

He destroyed his own leadership identity — the lighthearted internet dreamer who wanted to be liked. He destroyed the consensus culture he had spent years building. He destroyed the careers of executives he respected and the roles of co-founders he loved. He destroyed the organizational architecture that reflected his personality, because his personality was the thing most likely to kill the company.

At XPeng’s tenth anniversary in 2024, Xiaopeng summarized the journey with a phrase that has since become iconic in Chinese automotive circles: “在血水中游泳，一直游到海水变蓝” — swimming in blood water, swimming until the sea turns blue. The metaphor is not about surviving competition. It is about enduring the kind of suffering that changes what you are. The blood water was not the market. It was Xiaopeng himself — his instincts, his habits, his identity as a leader — and the sea did not turn blue until he had bled enough of himself away to become someone new.

He remains characteristically cautious. “We’re just walking out of the ICU,” he told reporters. “No company has secured a ticket yet.” But the man saying those words is not the man who entered the ICU. The chairman who wanted to be nice became the CEO who learned that leadership, at its most necessary, is the willingness to be disliked by the people you most want to impress.

Last updated: March 15, 2026