In one hundred and forty years, the auto industry has answered only one question

[[**One hundred and forty years of automotive history prove: once a supplier enters car manufacturing, it loses the trust of the entire industry. The coming-of-age ceremony for Chinese automakers must safeguard the three sovereignties of product definition, brand, and channels; only by keeping the platform layer neutral can a healthy ecosystem be built.**

**Key Points**

**1. A century-old iron law: suppliers do not build cars and do not compete with customers for profit**

Bosch has upheld the belief of "not building cars" for one hundred and forty years, becoming the global Tier 1 leader; Delphi, spun off by General Motors, went bankrupt because it served only its parent company and lacked market competition. This boundary is a law drawn with corporate life and death.

**2. The coming-of-age ceremony for Chinese automakers must master the "three-piece sovereignty package"**

Product definition rights, brand ownership, and channel leadership are all indispensable. In the joint-venture era, handing over these three rights in exchange for manufacturing capability led to an outflow of goodwill and profits; in the new energy era, any "deeply packaged" intermediate form of cooperation will collapse in a downturn because responsibility is diffused.

**3. Platform-layer neutrality is the load-bearing wall of the ecosystem**

In the era of smart electric vehicles, the platform layer—computing chips, autonomous driving algorithms, batteries, and so on—is highly concentrated. If these suppliers supply exclusively or build cars themselves, they will destroy industry-wide trust. CATL's choice not to build cars has proven this: restraint is the most expensive asset.

**4. Camps win battles, ecosystems win eras**

Vertical integration or exclusive binding splits the market into camps, creating strong short-term momentum but low long-term efficiency. A layered structure requires platforms to remain open, vehicle makers to safeguard sovereignty, and industrial belts to take root locally; only when all three stay in their proper places can shared prosperity be achieved. 2026-10-09 07:55 For one hundred and forty years, the auto industry has answered only one question Jinduan Quick View

This article is from the WeChat public account: Jinduan, author: Mu Yang

First, look at three scenes.

Scene one: 1886, Stuttgart. Robert Bosch opened a precision machinery workshop, starting with internal combustion engine ignition devices; his earliest customers were Daimler and Benz, a group of people who had just invented the automobile. Over the next one hundred and forty years, this company made the right choice on almost every multiple-choice question in the history of the auto industry.

Scene two: 1999, Detroit. General Motors shareholders approved one of the largest spin-offs in the history of American industry: stripping out its vast parts empire and forming an independent company called Delphi. On the day of the spin-off, Wall Street applauded, and management breathed a sigh of relief. No one expected that six years later, this company would enter bankruptcy court.

Scene three: today, China. Technology companies are pouring into the auto industry with intelligent capabilities. Between automakers and suppliers, between leaders and supporting players, a set of complex relationships unprecedented in this industry is evolving: deep cooperation, cross-sector equity stakes, platform binding, the transfer of sovereignty, and the recovery of sovereignty—all happening simultaneously every day.

Three scenes, one hundred and forty years apart, are in fact three askings of the same question: in the auto industry, who is qualified to sit in the driver's seat?

We have re-examined the history of global auto industry suppliers over the past one hundred and forty years and found that this boundary was not designed by anyone; it was drawn by the life and death of one company after another—Bosch proved one side with one hundred and forty years, and Delphi proved the other side with bankruptcy.

This article is divided into three parts:

History: how the Tier 1 landscape took shape and how the boundary was drawn.

Coming-of-age ceremony: where Chinese automakers stand on this boundary, and the questions they must answer.

First, explain a term unfamiliar to ordinary consumers but revered by everyone in the industry: Tier 1, first-tier supplier, meaning a supplier that directly delivers complete system assemblies to automakers.

Bosch, Continental, ZF, Denso, Magna—names that consumers can barely mention—constitute the true middle layer of power in the global auto industry. In an era when a fuel vehicle has more than twenty thousand parts and automakers' self-manufacturing ratio is increasingly low, it is Tier 1 that determines what a car's brakes, steering, transmission, and electronic control systems look like.

The starting point of this landscape is that Stuttgart workshop. Over the next one hundred and forty years, Bosch made the right choice on almost every multiple-choice question, but only one choice truly laid the foundation: never compete with customers.

Bosch can supply ESP electronic stability control systems to Mercedes-Benz, and it can also supply the same ESP to BMW, Volkswagen, and Toyota—because it does not build cars itself, it can take every company's order with a clear conscience. The R&D cost of the same technology is spread across the scale of the entire industry; this is the secret of Tier 1 profits and the dividend of its neutrality.

Note that not building cars has never been a "capability issue" for Bosch—given Bosch's technical depth and financial strength, building a car would not be fantasy. It is a matter of belief: precisely because it personally closed the door to becoming an automaker, it can stand at the table of all automakers at the same time.

General Motors made large-scale acquisitions of parts plants during its expansion era, bringing transmissions, spark plugs, and radiators into its own system; Ford was even more extreme, wanting to hold everything from iron ore to rubber plantations in its own hands—the steel, glass, and tire chain of the River Rouge industrial complex was the peak totem of the "vertical integration" faith. The logic sounded impeccable: build it yourself, costs are controllable, supply is secure.

What was the result? Its captive suppliers became a cost black hole: supplying only one parent company, unable to achieve scale, slow in technological iteration, unable to reduce costs, and forced to make unconditional concessions when the parent company bargained. By the end of the century, the two automakers had to tear down the empires they had built with their own hands: GM spun off its parts system as Delphi in 1999, and Ford followed in 2000 by splitting its system into Visteon.

Was splitting them off enough? No. Delphi filed for bankruptcy protection in 2005, and Visteon followed into bankruptcy proceedings in 2009. The legacy of a generation of vertical integration ended up in bankruptcy court. Meanwhile, the independent Tier 1 suppliers of the same era—Bosch, Continental, and ZF—became the mainstays of the auto industry. ZF even spent tens of billions of dollars to swallow TRW in 2015, then acquired Wabco in 2020, growing stronger and stronger.

Detroit proved one thing over seventy years: an automaker can own a parts plant, but it cannot own efficiency; a supplier can be kept, but being kept does not breed competitiveness. What captive divisions have always lacked is not investment but the test of market competition. A division with no customers to lose has no reason to cut costs or innovate.

(Later, some tried to overturn the verdict: wasn't BYD's vertical integration very successful? Note the key difference—BYD's FinDreams Battery sells externally, and its semiconductors supply externally; the internal divisions are always exposed to market competition. The success or failure of vertical integration lies not in "how much you do yourself" but in "whether the internal workshops must withstand the test of the market." Closed empires die of arrogance; open integration wins through clarity.)

3. Toyota's sense of proportion: however deep the blood ties, the books stay separate

The Japanese route offers a more refined version.

In 1949, Toyota spun off its electrical department to establish Nippon Denso, today's Denso. Toyota holds about a twenty percent stake, and the blood ties between the two are deep; but from the day of the spin-off, Denso operated independently and won customers independently, and its customer list long included GM, Ford, and a roster of European and American automakers. Aisin, another Toyota-affiliated supplier, has the same structure: equity ties by kinship, operations each standing on their own.

This "sense of proportion" was later written into every supply chain textbook: equity can connect, but brand and operations must be independent. Connection buys deep collaboration; independence buys industry-wide scale. Toyota's supplier system was once criticized as "closed," but even in the era when the criticism was harshest, Denso never sank to serving only Toyota. It was precisely this independence that let it survive the yen's appreciation and the bursting of the bubble, and grow into a leading global parts giant.

In 2022, in an era when power batteries accounted for thirty to forty percent of vehicle costs, the chairman of a leading automaker publicly complained at the World Power Battery Conference: automakers are working for battery makers. Public opinion was in an uproar, and calls along the industrial chain to "ditch CATL" echoed for a long time. With the bargaining power CATL had at the time, it had countless ways to reach downstream—taking stakes in automakers, binding channels, trading supply priority for a voice.

It did none of that. CATL continued to supply all automakers—including competitors locked in fierce mutual battles. Several years on, the "working for battery makers" talk has gradually faded, CATL has become the absolute number one in global power batteries (today, more than one of every three power batteries worldwide comes from its hands), and on top of its neutral supply, Chinese automakers have captured half of the global new energy vehicle market.

In the battery battle, Chinese industry won: it won because the supplier kept to the supplier's place. This was the first large-scale validation of an iron law a century old in the Chinese context—restraint is not weakness; restraint is the most expensive asset.

5. Three iron laws

Compress one hundred and forty years, and the global auto industry has drawn three iron laws in blood and tears:

First, a supplier does not compete with its customers for profit. Once a supplier reaches out to build vehicles, it becomes a threat to every customer, and orders from across the industry will leave it.

Second, the brand must belong to the automaker. What consumers remember is the badge on the hood; whoever holds the badge is the true master of the car. Cooperation in which the brand hangs in the supplier's hands has a landmine buried in it from the very start of property rights.

Third, deep collaboration and operational independence must coexist. Equity can connect, technology can be shared, but product definition, brand operation, and the channel system must belong to the automaker itself—otherwise it is not an automaker, just a licensed contract manufacturer.

Put thirty years of China's auto industry into this framework, and you see a structurally complete coming-of-age narrative: the student years, adolescence, the rite of passage. Three exams, but all testing the same question.

1. The adolescent trap: why the in-between state is dangerous

Chinese automakers are no strangers to the taste of "incomplete sovereignty." They first tasted it in the joint venture era.

In the joint venture wave around 1984, the framework was "market for technology": foreign partners contributed brands, product definition, and core technology; the Chinese side contributed plants, labor, and market. Thirty years on, manufacturing capability is indeed world-class; but review the property rights structure: brands were in foreign hands, product definition was in foreign hands, and the bulk of channel profits flowed along the joint venture terms back to the foreign parent countries. The Santana sold for decades and became a memory for a generation, but whose books that goodwill was recorded on goes without saying. The tuition paid in the student years was the long-term absence of three sovereignties.

In the ten years of new energy, the script turned a page, but the same structural temptation reappeared in new clothes: tech companies entered with intelligent capabilities, offering automakers "deep packaged" solutions spanning product definition, co-governed brands, and channel sales. In a favorable cycle, this was called complementary strengths—the tech company supplies the soul of intelligence, the automaker supplies the body of manufacturing. But hold it up to the iron laws of a century of history, and the structural risk of this "in-between state" becomes visible:

Each party holds only part of the property rights, so each party bears only part of the responsibility. Brand and product definition sit with one party, manufacturing and qualifications with another, and capital and ultimate decision-making power with the shareholders' meeting—when the industry enters a downcycle, every party can calculate the case for cutting losses, but none can calculate a reason to "save" it. For an automaker with full sovereignty, death is chronic: it can sell assets, raise financing, switch tracks; struggling itself is a right. For a "carmaker" with incomplete sovereignty, death is instantaneous: it depends on whether anyone in a shareholders' meeting a thousand miles away is willing to keep voting.

This is not a judgment about any one company, but about a type of structure. History has already taught this lesson twice—through Detroit's defeat and through the thirty years of the joint-venture era.

Figure: The three paths for tech companies crossing into the auto industry

2. Two proven paths: Xiaomi's $10 billion, Apple's turn

In the same era, tech companies voted with real money along these three paths.

The representative of the "all-in" path is Xiaomi. When it officially announced carmaking in 2021, it faced two "light" options: find an automaker to do contract manufacturing, or become a solution supplier. It chose neither—it built its own factory, obtained its own qualifications, used its own brand, built its own channels, and defined its own products. Lei Jun called it "the last major startup of my life," with a plan to invest $10 billion over ten years. Its first car launched in March 2024, with first-year deliveries exceeding 100,000 units; its second car launched in 2025 and continued to sell well.

With a check on the order of $10 billion, Xiaomi acknowledged one thing, which mirrors the Bosch family precept from a century ago: there is no asset-light shortcut to carmaking. Either go all in and become a true automaker, or stay out of the arena and be a true supplier.

The representative of the "retreat from the arena" path is Apple. In February 2024, Apple abandoned its carmaking plan after a decade of effort—the wealthiest and most patient team in consumer electronics history, with ten years and tens of billions of dollars invested, ultimately concluded that in the car business, entering means surrendering your entire position; half-hearted entry does not work. After the turn, Apple refocused its automotive business on a "supplier position"-style role.

More Chinese tech companies have long understood this: the drone giant spun off its automotive business into an independent intelligent driving solution company supplying the entire industry; the smart-driving algorithm company's client list spans central SOEs, new forces, and multinational giants; Horizon Robotics listed on the Hong Kong stock exchange in October 2024, becoming the "first smart-driving chip stock," and Volkswagen even invested about $2.4 billion to form a joint venture with it.

Both paths work. The only thing history has repeatedly disproven is the intermediate state. Bosch closed the door to carmaking from the outside; Xiaomi pushed the door open from the inside and moved itself entirely in. Different routes, same destination: both rejected the "one foot in" state.

3. Reverse output: from student to teacher

If the above answers "which side tech companies should stand on," another thread is shaking the industry even more: Chinese automakers are turning from students into teachers.

Thirty years ago, the report card of "market for technology" was indeed unsatisfactory—in the three major components of fuel vehicles: engine, transmission, and chassis, thirty years of pursuit never truly caught up. Switching lanes to new energy, the script turned the page.

In 2023, Volkswagen announced an investment of about $700 million in XPeng and plans to use XPeng's technology platform to develop Volkswagen's electric models in China—the first transaction in more than thirty years of China's auto joint ventures in which a multinational giant bought platform technology from a Chinese automaker. That same year, Stellantis spent 1.5 billion euros to acquire a 20% stake in Leapmotor and set up a joint venture to sell Leapmotor cars overseas through Stellantis's global channels; this Hangzhou automaker, which insists on "full-domain in-house development," subsequently rose to the top tier of new forces in deliveries.

Meanwhile, Chinese automakers' "in-house development list" keeps growing: NIO's in-house smart-driving chip is already installed in vehicles, XPeng's in-house chip has been taped out, BYD has pushed high-level smart driving down to 70,000-yuan-class models, sparking "smart driving equality," and the cost curve of intelligence has been rewritten by Chinese automakers themselves. In 2024, BYD alone sold more than four million new energy vehicles—on any table in the world, that is a number that rewrites the industry landscape.

From "market for technology" to "technology for market," in thirty years Chinese automakers have completed the identity swap. This is the first half of the coming-of-age ceremony: adulthood in capability.

4. The measure of the coming-of-age ceremony: the three-piece set

But adulthood in capability does not equal adulthood in status. The measure of this coming-of-age ceremony must shift from "can build" to "who decides."

Our observation is that the sign of an automaker truly coming of age is the sovereign ownership of three things:

Product definition rights: who decides what this car looks like, who it is built for, and what its selling points are; brand ownership: whether the trademark on the hood is owned by us, and whether the accumulated goodwill belongs to us; channel leadership: in whose stores the cars are sold, in whose systems the user data sits, and who holds pricing power.

These three are exactly the Chinese version of the three iron laws from the previous article. Measure the joint-venture era with this ruler: the three sovereignties were handed over for thirty years in exchange for manufacturing capability, at the cost of long-term outflow of goodwill and profits. Measure some "deep cross-sector collaborations" still underway today: no matter how impressive current sales are, as long as the three rights are incomplete, structurally it is still adolescence.

Figure: The "three-piece sovereignty set" of an automaker's coming-of-age ceremony

I must add a sentence here immediately, lest it be misread: sovereignty resting with us does not mean doing everything ourselves. This is where the coming-of-age ceremony is most easily led astray. The lessons of vertical integration are still lying in Delphi's bankruptcy filings; if "full-stack in-house development" slides into "building everything ourselves," that is an infantile disorder, not a coming-of-age ceremony.

The adult-style division of labor should be: the three-piece sovereignty set yields not an inch, and for everything else, buy what should be bought and cooperate where cooperation is due. Buying Bosch's ESP is no disgrace, using CATL's batteries is no disgrace; what is disgraceful is having someone else decide your product definition, having your brand under someone else's name, and having your users held in someone else's channels.

Measure today's Chinese automakers with this ruler: among the top few, the three-piece sovereignty set is basically complete; in the middle and tail ranks, a considerable number are still "borrowing a boat"—the boat can be borrowed for a while, but the price of the ticket will sooner or later be repriced at some shareholders' meeting.

Pull the camera back, and you will find that the industrial power structure of the intelligent electric era is growing into a new shape.

In the fuel-vehicle era, the Tier 1 landscape was "multiple strong players side by side": Bosch handled electronic control, ZF handled transmissions, and Continental handled chassis, each guarding a domain of expertise. In the intelligent electric era, value is highly concentrated toward "computing + software," and a new layered structure is emerging—

English text is being prepared.

Vehicle layer (vehicle brands and manufacturing): dozens of automakers are engaged in close-quarters combat in the most brutally competitive market.

Figure: The industrial layering structure of the intelligent electric era

The essence of this layering is: the foundation of intelligence concentrates toward a few platforms, while vehicles and brands disperse across many automakers. It is remarkably like the "Windows-Intel and PC makers" of the PC era, and the "Qualcomm-Android and phone brands" of the smartphone era—history always repeats itself in a different outfit.

And whether this structure is healthy, judging from a century of history, depends on only one thing: whether the platform layer remains open and neutral.

2. Neutrality is the most expensive asset

The neutrality of the platform layer is the load-bearing wall of the entire ecosystem.

Bosch's century and CATL's recent years tell the same truth: a supplier treating the whole industry alike is its most valuable asset, and more valuable than any single technology. Because customers entrust the "soul" of intelligence (chips, algorithms, data architecture) to you, and the precondition of that trust is confirming that you will not turn from guest into host: that you will not use your position as a supplier to take away their voice, that you will not leave a back door for "your own people," and that you will not one day enter the game yourself and become a competitor.

This also means the new Tier 1s must answer some old questions: when an automaker wants to use orders to trade for exclusivity, or equity stakes to trade for priority, do you accept or not? New structures of automakers taking stakes in intelligent platforms have recently appeared in the market—deep binding and conflicts of interest are separated by a fine line; Bosch's answer was "equity can be discussed, operations are not meddled with," and this sense of proportion is worth framing on everyone's wall. When the platform's own technology roadmap conflicts with a major customer's short-term interests, whose voice wins? When the industry winter comes and orders shrink, will you be unable to resist entering the game yourself, trading a supply-chain position for a ticket to the vehicle race?

Every temptation to "enter the game" is an overdraft on neutrality, the most valuable card in hand. Detroit's vertically integrated empire and the exclusive bindings in consumer electronics all ruined the same thing: they cut the market into inside-the-wall and outside-the-wall. And the source of efficiency in China's supply chain over these thirty years has been precisely openness and flow. This lesson does not belong to the auto industry; it belongs to all of Chinese industry.

3. The arithmetic of industrial belts: where the brand stays, the profit stays

Shift your gaze from equity structure to the geographic map, and the boundary question lands on something more concrete: industrial belts.

The map of China's automotive industrial belts over the past thirty years was basically "a distribution map of joint-venture factories": Shanghai, Changchun, Wuhan, Guangzhou—wherever capacity landed, the industrial belt grew there—but the brands belonged to others, the bulk of profits flowed to others' headquarters, and the industrial belts long remained suspended in a state of "manufacturing locally, branding externally."

In the ten years of new energy, this map has been redrawn, and its nature has changed.

Figure: The "sovereignty arithmetic" of China's automotive industrial belts

This is the significance of the "coming-of-age ceremony" at the regional-economy level: when product definition power, brand ownership, and channel leadership return to the automakers' hands, the arithmetic of the industrial belt changes completely—

The brand stays on the industrial belt, and the compounding of goodwill stays local: a valuable brand means long-term accumulation of headquarters functions, R&D jobs, tax sources, and talent, which no contract factory can match.

Profit stays in the chain, and the ammunition for reinvestment stays local: profit at the vehicle-assembly stage stays in the region, seeps down the supply chain, and becomes orders for tier-2 and tier-3 suppliers, technical upgrades of production lines, and workers' wages.

The certainty of orders stays local: manufacturing for someone else's brand, a factory always faces the "relocation option"; manufacturing for one's own brand, home ground is home ground.

Conversely, if sovereignty flows out—the brand hangs under a partner's name, channels are held by others, pricing power flows out with revenue-sharing agreements—then no matter how many super-factories are built on the industrial belt, they are merely prettier contract-manufacturing workshops. The upgrading of an industrial belt is in essence not an upgrading of capacity, but the return of sovereignty.

3. A good ecosystem doesn't require taking sides

Pulling these threads together, we can set one test for a "good industrial ecosystem," and only one: do the companies within it need to "take sides"?

If a platform supplies only "its own people," automakers can only ask, "Whose coattails should I cling to?" If a giant locks in partners with exclusive terms, suppliers can only ask, "Can I still work for anyone else?" The vertically integrated empire and the repeatedly discussed "choose one of two" phenomenon are both bad for the same reason: they carve the market into camps and turn cooperation into dependence.

China's auto industry is standing at this fork right now. On one side is campization—platforms and vehicle makers deeply bound, cross-shareholdings, exclusive exclusivity; in the short term every pairing looks formidably powerful. On the other side is layering—platforms open and neutral to the whole industry, vehicle makers with complete independent sovereignty, linked by contract rather than "personal dependence." The verdict of a hundred years of history is already written: camps win battles, ecosystems win eras. Detroit's camps lost to Toyota's ecosystem; the feature-phone camps lost to the smartphone ecosystem.

For China, a country with the world's most complete automotive supply chain, the answer should be even more self-evident: platforms hold neutrality, vehicle makers hold sovereignty, industrial belts hold the land they are rooted in—only when all three keep their proper places is there true shared prosperity.

History says: a supplier's boundary is not morality but law—Bosch's hundred years and Delphi's bankruptcy are the two sides of the same iron rule;

Automakers say: the measure of coming of age is not sales volume but three sovereignties—product definition, brand, and channels; lacking any one, a company is structurally in its adolescence;

Ecosystem says: a good industry doesn't require taking sides—platform neutrality, vehicle-maker autonomy, industrial-belt rootedness; only when all three keep their proper places is there shared prosperity.

What is beyond doubt is that the "era of layering" has its own new risks: the concentration of the platform layer may be brewing the next kind of "hidden power." Today's compute chips are to smart cars much as the engine once was to fuel vehicles; Bosch-style giants are already showing signs, and the scene of automakers queuing for chips was familiar during the global chip shortage a few years ago; and the data-and-algorithm closed loop of intelligent driving may let "first-mover platforms" exert generational suppression over "late-arriving automakers." A hundred years ago, automakers learned through bloody lessons to draw boundaries with Tier 1s; a hundred years later, that boundary may need to be drawn once again.

But no matter how the boundary moves, one thing has not changed in a hundred and forty years, and probably won't change in the next thirty—

The steering wheel can be handed to algorithms, but the driver's seat must belong to yourself.

An automaker may not know the technological answer for the next decade, but it must always know where it is sitting. This is the same conclusion jointly verified from four directions—by Bosch with its family precepts, by Delphi with bankruptcy, by Apple with a pivot, and by Xiaomi with ten billion US dollars.

It is also the one answer that China's auto industry, from student, to youth, to adulthood, cannot lose in this long coming-of-age ceremony. Channel: Cars and Mobility

431 articles published on Huxiu. This content is published with the author's authorization; the views represent the author only and do not represent Huxiu's position. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]