Pioneers Insight Method Research Author
Supplier Behind Jiyue's Implosion: Ambition and Public Trust Crisis
Back to Episodes

Supplier Behind Jiyue's Implosion: Ambition and Public Trust Crisis

Summary

  • Supplier pre-funding is the default unwritten rule for new-energy vehicle brands buying traffic on Douyin, and the risk-reward is terrible: roughly 1% in rebate margin for tens of millions of RMB in funding exposure. 李红星’s breakdown is straightforward: Douyin requires payment upfront, while automakers make pre-funding a condition in tenders. “You can choose not to pre-fund and submit a bid, but you won’t win.” The result of industry-wide competition is that suppliers hand over their lifeline first, then are afraid to press clients for payment.
  • 李红星’s RMB37M in pre-funding for Jiyue followed a textbook “boiling frog” path: he began operating its livestream channel in May 2024 and was paid monthly, then won an RMB18M annual traffic-buying contract in August and funded nearly RMB30M over the next 3 months. Jiyue’s 07 launch pushed monthly traffic spend from around RMB2M to more than RMB10M, well beyond the contract. His recommendation to “bring in another supplier; this is too aggressive” was ignored. After RMB5M failed to arrive in October, pressure from the client—“Do you want to keep doing this? If not, we’ll find another supplier”—and wishful thinking led him to fund another RMB5M in November, forcing him to mortgage his home and borrow more than RMB20M from close relatives.
  • His complete reversal on the self-imposed rule of working only with new-energy brands is the episode’s most revealing admission of error. His original logic was that EV startups thought more like internet companies, focused on results and less prone to relationship-driven interference. His conclusion now: “I was wrong.” Traditional automakers “might at most leave you with no profit; they won’t take your life, but new-energy automakers can blow up at any time.” He now hears that some automakers will accept prepayment, but the price is that suppliers must give up every bit of profit and merely run volume through their books.
  • 李红星 attributes Jiyue’s collapse to 3 layers—its ownership structure, Baidu, and 夏一平—and rejects a single-cause explanation. His analogy: 2 parents with their own children and households adopted a child, then hired a nanny to look after it from outside; no one truly cared. Baidu, in his view, “couldn’t make anything work, put money in, saw it wasn’t working, and rushed to cut its losses.” 夏一平 had never run a company as its top executive, lacked financial controls, and lost management discipline by going all-in on marketing.
  • 卫诗婕’s central pushback is that marketing was never Jiyue’s way out; 夏一平’s biggest failure was consciously or subconsciously avoiding the financing problem. 李书福 arranged a meeting with 李彦宏 without taking him, yet he remained optimistic afterward that “we can just keep selling cars.” In the documentary, he even says financing is the CFO’s job. But she also concedes that amid the 2024 price war and the 2 major shareholders’ divided attention, “Could a more capable, more experienced CEO really have solved this?” 李红星 fully agrees.
  • The shared conclusion about a “public trust crisis” is the episode’s real endpoint: even a marquee project backed by Baidu and Geely could quietly stop paying social insurance for more than 5,000 employees and deliberately delay supplier payments, “and then the entire business world will lose trust.” 李红星 says that if the outcome was already beyond saving, the company should have faced it at lower cost. 卫诗婕 compares the major shareholders’ handling to adults fighting, flipping the table, and leaving others to clean up. Large companies should make a clear decision before collapse: put up more capital and continue, or “give the public a dignified, responsible way to end.”
  • The irony is that the operating data show this was not a worthless hand: after 李红星 took over, lead costs fell from more than RMB400 to more than RMB200, monthly leads reached 10K-20K, market share was described by him as 7%-plus and 8% higher, and half of the sales growth came from livestream leads. Engineers pleaded, “Give us 10 more days and the 3.0 system will be released,” while owners said referrals could double sales—“This was a hand we could keep playing.” To this day, Jiyue has made no official contact with suppliers; employees even went to his location late at night to try to retrieve livestream vehicles he was holding, and the police have opened a case.

Deep dive

1. Laughter at the Screening: “Why Is This Person Sitting in the CEO’s Chair?”

  • The opening comes from a March 21 screening of Tencent’s documentary Torrent 2, hosted by 卫诗婕. When 夏一平 said the Jiyue case could one day be written up as a classic business case, the room broke into laughter. She stopped to ask what, exactly, people were feeling when they laughed.
  • The sharpest comment came from an automotive advertising executive: “The more I watch this documentary, the more confused I become. Why is this person sitting in the CEO’s chair at a new-energy vehicle company?” 夏一平 could frame grand narratives but did not understand cars at the level of details. Compared with XPeng bringing in 王凤英, 李斌’s self-criticism, and 雷军’s approach to selling cars, “this may have been a flawed premise from the start.” Even when the team was doing Douyin, “there were simply no cars to deliver.”
  • The context is stark. When Jiyue ground to a halt in December 2024, nearly 1,000 employees, owners, and suppliers surrounded the company. More than 5,000 employees discovered that social-insurance contributions had stopped months earlier; everyone had been kept in the dark until the company suddenly imploded.

2. 李红星: “Two Years of Easy Wins” in Douyin Livestream Growth

  • 李红星 describes himself as the owner of Sina Culture and a supplier harmed by Jiyue, with 19 years in the internet industry. He previously worked with Google China as an agency, ran ChinaByte, and from 2016 worked at ByteDance on automotive business, Douyin user growth, and local services. He left at age 40 in 2022 to start a Douyin advertising agency.
  • His demystifying review of Douyin’s livestream rise is blunt: it was not primarily the result of superior creator operations. “Do you know what the core reason was? The pandemic.” His team was created in 2020 to drive livestream growth, just as people were forced to broadcast from home. They “got 2 years of easy wins,” with performance growing 700%, 800%, and 900% each time. Even the light show that illuminated the Qiantang River was “icing on the cake because the whole business was taking off.” “When you choose the right thing, it becomes very easy.”

3. The Agency Business: RMB300M in Revenue and “Hard-Earned Money” at 1%

  • In his first year as an entrepreneur, revenue exceeded RMB300M, while profit was “less than 1%, basically around 1%”—roughly RMB3M. That was more than enough for a small team, but most of the profit came from platform rebate incentives.
  • What justified that 1%? The agency took on account opening, finance, and media-buying operations for the platform. “If every client signed directly with Douyin, how enormous would its finance team have to be?” 卫诗婕 summarized the business as “hard-earned money,” and 李红星 agreed: “That’s more or less what it is.”

4. 3 Days and 3 Nights at the Collapse Site: A Marketplace of a Headquarters and a Supplier Queue That Never Moved

  • The day after the blowup, 李红星 went to Jiyue’s Shanghai headquarters. Police cars filled the street and officers surrounded the company. He borrowed an employee access card to get inside that afternoon. “A company with thousands of people that had once been orderly had turned into a marketplace.” With tens of millions of RMB trapped inside, the odds of recovery were poor, and “the sense of loss is hard to imagine.” He compared it with a farmer watching bandits carry away an entire season’s harvest: shouting does nothing.
  • The groups inside had very different leverage. Employees were “very orderly”—in quotation marks—chanting slogans and taking turns sending representatives to negotiate. Owners also organized. Suppliers numbered only around 10, describing themselves as representatives and asking for a dialogue, but “nobody paid attention to you.” To this day, no official from Jiyue has come out to engage with suppliers.
  • 夏一平 was actually there throughout, being rotated through negotiations in a meeting room of only a dozen or 20 square meters. “The moment he fell asleep, someone would push him awake: you can’t sleep.” Security escorted him for food deliveries and bathroom breaks, and on the first day he managed only 1 hour of sleep near dawn. 李红星 stayed for 3 days and 3 nights without ever getting a place in the queue—“Maybe staying inside made you feel there was still some hope.”

5. From Anger to Postmortem: The Ownership Structure as “An Adopted Child with a Nanny”

  • At the time, 李红星’s emotion was hatred: “This situation was caused by him alone.” He says he was unable to think clearly then. Looking back, however, “this did not happen because of any one person; it was a systemic event.”
  • His 3-layer attribution is worth preserving in full. First, the ownership structure was defective from the beginning: “2 parents who already had their own children and their own households adopted a child and hired a nanny outside to look after it—who would genuinely care about this adopted child?” Second, Baidu had its familiar pattern: “Everyone knows what it’s like… no strategic conviction, unable to make anything work, putting money in, deciding it wasn’t working, and rushing to cut its losses,” while its own revenue declined and it fell from the BAT tier. Third, 夏一平 had never been a top executive, lacked financial controls and management experience, and “if he spent so much energy on marketing, he would inevitably neglect other areas.”
  • One piece of background: Baidu could not obtain a license to manufacture cars, and Geely contributed a sub-brand qualification, which made Jiyue possible. 李红星 says he “roughly understood” that history at the time but “never imagined it would lead to such a major conflict between the 2 companies.”

6. How RMB37M Was Funded: From Prepaid Livestream Operations to Pre-Funded Traffic Buying

  • The timeline matters. Starting in May 2024, he operated Jiyue’s official livestream channel. His initial requirement was prepayment: “We don’t do it without prepayment.” Jiyue did not prepay, but it paid on time. The major funding exposure came in August, September, and October, when he put up nearly RMB30M; the cumulative amount reached RMB37M.
  • In August, Jiyue decided to hire a traffic-buying company. 3 agencies bid, and he won on expertise and price by returning 2.5 percentage points from the platform rebate to the client—“sacrificing profit.” The tender amount was RMB16M for 1 year, later signed as an RMB18M annual framework. His assessment at the time was that monthly spending would be only RMB1M-2M, with manageable and coverable risk.
  • On rumors that he had paid kickbacks, he was unusually emphatic: “I didn’t give 夏一平 a single cent, not a penny of benefit, and I didn’t give anyone at Jiyue even the slightest benefit. That’s something I can say with my head held high… If I took my tiny bit of profit and then gave it back to them, I’d really have to be out of my mind.”

7. The Difference Between Livestream Operations, Traffic Buying, and an Integrated Model

  • Livestream operations means executing the broadcast: recruiting and managing hosts, setting broadcast times, arranging lighting and the set, and writing scripts. Traffic buying means “using money to buy traffic”—paying to bring users into the livestream.
  • Established agencies could make enough from traffic buying alone, while newer agencies needed a way to break through. 李红星 therefore built an integrated “broadcast-plus-traffic” model, putting livestreaming and media buying under 1 team so “there would be no mutual buck-passing or shifting of responsibility.” After the Jiyue case worked, the platform invited him repeatedly to share the model and recommended him to other brands. Only a dozen or so official agencies nationwide can handle automotive advertising on Douyin, leaving clients to choose from that limited group.

8. His “New-Energy Only” Rule Was Wrong

  • When he started the company, 李红星 decided to work only with new-energy brands. Traditional automakers offered lower margins, were harder to work with, had longer payment cycles, more rigid thinking, harsher pressure on suppliers, and more complicated relationships. EV startups seemed more internet-minded, more results-oriented, and easier to communicate with.
  • 卫诗婕 asked whether that also meant rent-seeking and kickbacks were more common at traditional automakers. 李红星 said he had not encountered anything comparable at EV startups, but later admitted directly: “Looking at it now, I was wrong.” In terms of safety and operating discipline, “no matter how a traditional automaker delays your payment for 6 months or a year, it will ultimately pay you. At worst, you don’t make money; it won’t take your life. But new-energy automakers can blow up at any time.”

9. Why HiPhi’s Precedent Did Not Stop Him: “I Trusted the 2 Major Shareholders Too Much”

  • Asked about earlier cases in which suppliers failed to recover pre-funding, he immediately said, “HiPhi—didn’t HiPhi blow up?” But his mindset at the time was that “it wouldn’t happen to me.” Baidu and Geely were building Jiyue together; “Baidu was, in any case, one of the top companies,” and he had heard that the company had more than RMB1B in cash and should not face problems in the short term. 卫诗婕 noted that later reporting showed Jiyue had only 2 months of cash left as early as April 2023. He had no idea beforehand.
  • His self-criticism is unsparing: “Blindly trusting so-called big companies, blindly confident, lacking judgment… Just because others have done something, or because everyone is doing it, doesn’t mean it’s right. From a business perspective, using so much capital to earn such a low return is not a wise decision, whether you look at it as an investment or as a business.” The head of the core business had warned him early: “This business is quite dangerous. After doing it for 6 months, stop pre-funding.” They “were aware of the risk, but never imagined it would arrive so quickly.”

10. The Uncontrolled Fall: Funding RMB5M Even After He Knew It Probably Would Not Be Repaid

  • The turning point was the 07 launch. In June and July, monthly traffic spend for the model was around RMB2M. From August and September, new-car launches pushed monthly spend close to RMB10M, far beyond the contract. 李红星 explicitly said, “The entire traffic plan is far beyond the contract we signed. I recommend hiring another supplier; this shouldn’t be so aggressive.” But “they were the actual decision-makers, and they wouldn’t listen to you.”
  • The RMB5M due in October did not arrive. He began pressing for payment and understood the problem by the end of the month. In November, even knowing that “there was a high probability the money would not come back,” he funded another RMB5M. “That RMB5M really should not have been funded; I was holding on to a bit of wishful thinking.” Procurement and business staff kept saying payment was “about to arrive,” and even pressured him in the opposite direction: “Do you want to keep doing this? If not, we’ll find another supplier. You can’t even pre-fund this little bit of money?” He stopped the livestream in late November and gradually halted traffic buying.
  • He lays out the supplier’s structural vulnerability clearly: “You hand over your lifeline to someone else first… If you stop everything for him, he may really say, ‘Why should I settle your payment? I’ve found a new supplier.’ So you can only fund part of it first, and get part of it back.” By September and October, after funding more than RMB20M and starting to borrow from relatives and friends, “the risk had become completely uncontrollable.” He mortgaged his home to keep going.

11. Douyin’s Rules: Pre-Funding as the Admission Ticket, Fast-Moving Consumer Goods as a Closed Loop and Autos as a Long Chain

  • The unwritten rule works as follows: Douyin traffic buying requires suppliers to prepay the platform, and clients make pre-funding a condition in tenders. “Suppliers A, B, and C are bidding. You can choose not to pre-fund and submit a bid, but you won’t win. That’s the result of industry-wide competition.” Someone has to fund the gap, and “who is that someone? The supplier.” 李红星 acknowledges the transaction is explicit: “This is a business you chose.”
  • Fast-moving consumer goods do not require the same funding because transactions close online: the amount spent and the amount returned can be calculated. Cars are sold through leads, with offline conversion and a longer chain, so the business follows traditional advertising economics; the margin on pre-funding is “somewhat more attractive.” What is the point of FMCG campaigns that break even? Year-end Douyin bonuses and rankings. “People would rather not make money as long as they can win the name—this is the default unwritten rule every Douyin agency is fighting over.”
  • One marginal change after the blowup is that 李红星 now hears some automakers are willing to accept prepayment. “Suppliers aren’t as stupid as they used to be.” The trade-off is that there is “basically no profit at all; they make you spit everything back out. It’s just running volume through the books.” Some companies still accept those terms for case studies and financial-statement revenue.

12. The Data: Lead Costs Halved and Market Share Near 8%, but ROI Could Not Be Calculated

  • After 李红星 took over, monthly leads exceeded 10K starting in July and August, reaching 17K-20K at the peak, with sales conversion around 2%. He described market share as “7%-plus, 8% higher,” leaving the exact metric ambiguous. “Half of Jiyue’s sales growth came from customers generated by livestream leads.” The previous supplier’s performance was poor and failed to meet lead targets; lead costs were above RMB400, while he brought them down to above RMB200. Jiyue told him its conversion rate was “basically 30%-40%, sometimes 40%-50%.” After the 07 launch, leads and sales orders were 2-3x those for the 01: the 01 sold 500-800 units, while the 07 sold 1K-2K.
  • The real ROI, however, was impossible to determine. Jiyue kept the actual number of cars sold through Douyin “confidential internally; they wouldn’t tell us.” 卫诗婕 pointed out the absurdity: “The normal logic is that they should give you the number and audit whether your ROI met the target. How could they keep it confidential?” He could assess only individual lead cost and market share.

13. Jiyue’s 3-Stage Marketing Leap: Tesla, NIO, and Finally All-In on Douyin

  • 李红星 saw the marketing strategy evolve firsthand. First came “benchmarking Tesla: we won’t spend on advertising; we have complete confidence in the product”—a form of “overconfidence” directly tied to the poor sales of the 01. Then Jiyue tried NIO-style owner referrals and word-of-mouth, “which also failed.” In 2024, investors demanded volume from 夏一平. “He thought and thought: where was the problem? Marketing—that was probably the best answer he could give.” The company then went all-in on Douyin livestreaming. 李红星 suspects a consulting team helped design the livestream and platform strategy, while 夏一平’s personal account was run by that team.
  • He offers a qualified defense of going all-in on Douyin. Under the traditional model, “you don’t see any splash from a new model launch unless you throw several billion RMB at it.” Concentrating the budget on Douyin, where the traffic was, was “from one perspective, a fairly reasonable solution at the time.” Monthly spending above RMB10M during the launch period was not excessive compared with the tens of millions of RMB that new-car launches typically consume.

14. 卫诗婕’s Pushback: Marketing Was Not the Way Out; Avoiding Financing Was the Core Failure

  • Her rebuttal deserves to be preserved in full. 夏一平’s biggest failure as CEO was “consciously or unconsciously avoiding the question of financing.” There was no formal, effective communication with 李彦宏 or 李书福, even though that was “something only the CEO should do and only the CEO could do.” His answer was to sell more cars first, then gain the confidence to ask shareholders for money. The most revealing detail: he asked 李书福 to arrange a meeting with Robin, but the 2 executives met directly “without taking him.” He remained optimistic internally afterward: “There’s no major problem; we can just keep selling cars.” 卫诗婕 says she does not know whether the senior executives conveyed contradictory signals or whether he was simply giving himself a reason to keep going.
  • She was also unsparing about the documentary itself. Watching it made her “not particularly comfortable,” because it seemed in places to excuse 夏一平. “He could actually say that financing was the CFO’s job. None of this is something a mature CEO should say.” But she also asked the harder question: with the 2024 price war intensifying, Baidu losing interest in building cars, and Geely still selling many of its own brands, “Could a more capable, more experienced CEO really have solved this?” 李红星 said he “particularly agreed” with her view: sales were an operating problem, while the CEO’s job was to solve the financial problem and discuss the funding shortfall with Robin and 李书福.

15. The Public Trust Crisis: “They Flipped the Table Like Children Playing House, Then Left Others to Clean Up”

  • 卫诗婕 defines the public significance of the episode this way: “If a marquee project backed by 2 nationally recognized giants can act outside the rules, quietly stop paying employee social insurance, and deliberately delay supplier payments, then the entire business world will lose trust. That is the truly damaging impact of this event.” If death was inevitable, the company should have “met that inevitable death at the lowest possible cost.”
  • She compares the major shareholders’ handling to children playing house: several adults fight, say “we’re not playing anymore,” flip the table, and walk away, leaving others to clean up and compensate employees and cover social insurance. The responsible approach would have been to put up more capital if they intended to continue; if not, they should have “given the public and the owners a dignified, responsible way to end.”
  • 李红星 strongly agrees with her assessment of the CEO’s responsibility. 夏一平 pinned the last lifeline on sales, which was an operating issue; the CEO needed to address the financial issue and discuss the funding gap with Robin and 李书福. He even believes the position was not hopeless: “You can’t call it an absolute royal flush, but it was a hand we could keep playing.” With more time and Baidu diluting its stake to let new capital enter, “I think there were some possible solutions, but the situation behind it was complicated, and we don’t know.”

16. Why Suppliers Stayed Silent, Why He Spoke, and How to Move Forward

  • 卫诗婕 explains why suppliers made far less noise than employees. Employees shared the same predicament and could generate sympathy; suppliers each had their own calculations. Many still had business with Geely and other brands and “didn’t want to speak out for fear of affecting other business.” When a large company has multiple shareholders, the losses are divided and “don’t hurt as much.” Small companies often go quiet after defending themselves for a while because “they really are the weaker group and don’t know what to do.”
  • His reason for coming forward was direct: “I’m an individual business owner, and all the money was secured against my own home—that is the most direct connection. And I believe that when something is right, you have to stand up and say so.” Even a small lesson for suppliers would be worthwhile.
  • His account of recovery is carefully hedged: “I can’t say I’ve completely moved on; at minimum, I’m still facing life actively.” He has 2 rules. First, “don’t rush to say I must turn things around immediately or struggle desperately.” Second, exercise and reading: he runs for 1 hour after 6 a.m., reads at the office, and reads again at night. He reads Munger and Dalio because successful people may have gone through mistakes and hardship far worse than his own. Dalio’s personal liabilities “were not simply tens of millions or hundreds of millions of RMB,” yet he still climbed back step by step. He cites what Dalio told 王岐山: “If every conflict could be resolved before it became a sharp problem, there would be no heroes in this world. Call it motivational fluff or whatever you want—that is simply the fact.”
  • His largest creditor is the close relative who lent him more than RMB20M and has still not demanded repayment. “That is an important reason I’ve been able to stay relatively optimistic and recover quickly.” He receives no understanding from some employees affected by the team’s downsizing, including one pregnant employee whom he is still paying. The debt case is being litigated in Hangzhou. “The process will be very long… what the outcome will be is not something you can control.”

17. The Signature Wall, “Give Us 10 More Days,” and the Absurd Late-Night Finale

  • In front of the signature wall saying that “the name of every Jiyue employee will surely remain in the history of China’s smart-auto industry,” 李红星 asked: “How did a perfectly good wall become this? How did a perfectly good company become what it is now?” The engineers’ reluctance to let go hit hardest: “Give us 10 more days. The 3.0 system will be released. We only need 10 days; we’ll work day and night”—but “the day we were waiting for never came.” An owner showed him a conversation between his daughter and the car’s voice assistant; the transcript alternates between SIMO and CMO. Another owner said, “If it had survived just 1 more month, sales should have doubled again. Referrals to 3 or 4 people around us had already turned into deposits.” The business contact who worked with him was still saying the day before the collapse, “Boss Li, this company won’t have a problem. Keep going”—“They couldn’t believe it themselves.”
  • Around 11:30 p.m. one night recently, 3 or 4 Jiyue employees quietly came to unlock and drive away the livestream vehicle left at his location. He discovered them, called the police, and a case has been opened. His legal position is clear: the vehicle-loan agreement had not expired, and Jiyue still owed tens of millions of RMB. “This is property held under a lien, and I have a lien right. You come and steal the vehicle for your own use—isn’t that ridiculous?” After police contacted them, the employees produced identification confirming that they worked for Jiyue and said “the company had arranged for them to collect it.”