Skip to content

Strategie & Business

Spider-Man in the Cockpit: When Success Is Only Administered

BMW is playing film promos on the screens of cars that cost 100,000 euros. Two weeks earlier, its automotive margin fell to 2.3 percent. Both are the same story. With an interactive timeline from 2020 to today.

Jonas Höttler10 min read
Contents9 sections

What happens when a company has been right for thirty years.

In late July 2026, people start their cars and get cinema. A banner appears on the control display. Tap it and you get 19 seconds of Spider-Man, with music and matching ambient lighting. It runs in more than 70 markets, on every BMW built from July 2020 on with operating system 7 or newer. Some of those cars cost $160,000.

The reaction is not ambiguous. Forums and social media produce words like "dystopian" and "tasteless breach of style". One sentence keeps coming back: this needs to be made illegal before everyone starts doing it.

BMW explains that this is not classic advertising but a broader brand partnership. The animation is not meant to promote the film. It is only visible if the driver taps it. In-vehicle advertising is not planned for the future either.

That last claim is the most interesting thing about the whole episode. Because in late 2023, BMW infotainment chief Stephan Durach told reporters: “Ultimately, in my view, the car is the last private retreat.” Selling the screen to play advertising on it was something he could not imagine.

Two and a half years later, Spider-Man is running.

The question is not why BMW changed its mind. The question is what happened between late 2023 and July 2026.

The numbers nobody plays into the cockpit

Eleven days before the animation started, BMW published its half-year results. They are the actual caption to this picture.

In the second quarter of 2026, the EBIT margin of the Automobile segment was 2.3 percent. A year earlier it was 5.4 percent. Group pre-tax earnings fell 35.1 percent to €1,697 million. Revenue dropped 7.9 percent to €31.3 billion. For the full year, BMW expects an automotive margin of 1 to 3 percent.

One percent margin. At a manufacturer whose strategic target corridor read 8 to 10 percent for years.

The driver is China. BMW delivered 261,800 vehicles there in the first half, down 20.4 percent. In the second quarter alone the drop was 30.2 percent. Europe is the group's largest sales region again for the first time since 2013. Not because Europe grew. Because China collapsed.

And the reason is not cyclical. In China, the retail penetration of new energy vehicles passed 60 percent for the first time in April 2026, at 61.4 percent. BMW's xEV share in the Chinese market was 9.8 percent in the second quarter.

That is not a dip. That is a market that moved on without the company.

Drag the slider

The timeline below stacks three things that sit in separate chapters of an annual report: what the product earns, where it earned it, and what the customer gets to see of the saving.

Every number comes from BMW publications. No estimates, no model.

INTERACTIVE

The decline, year by year

Drag the slider. On the left, the margin the product earns. On the right, what the customer gets to see of the saving.

0%4%8%20212223242526
Automobile EBIT marginChina share of global deliveries
Automobile EBIT margin2.7 %Pandemic year
Deliveries in China777,379+7.4 %
China share of global deliveries33.4 %
Margin index (2021 = 100)26-74 vs. 2021
Decided on the inside

Cost discipline in crisis mode. Free cash flow still beats the prior year at €3.4bn.

Visible on the outside · Product signal

Nothing. The customer sees none of it, because the saving happens where they do not sit.

All figures from BMW Group publications. 2026 shows half-year numbers; the second quarter alone came in at a 2.3% margin.

Read in sequence, it is a choreography.

2021 is the year everything works. Margin 10.3 percent, a China record of 846,237 vehicles, every third car sold goes there. No cost programme, a raised dividend.

2022 is the first year China falls. In the same year, seat heating appears in the ConnectedDrive store: €17 a month, €385 one-off, for hardware already built into the car. The customer bought it. BMW locked it.

2023 is the year of reversal and of the promise. The hardware subscription is withdrawn for lack of acceptance. And the car is declared the last private retreat.

2024 and 2025 halve the margin. 6.3 percent, then 5.3 percent. China loses roughly 200,000 vehicles in two years.

2026 brings both together: 8,000 jobs by end of 2027, around one billion euros of special budget purely to restructure administration, then roughly one billion euros of savings per year from 2028. And Spider-Man on the display.

Why this is not saving

There are two ways to repair a margin. One takes time: better products, shorter development cycles, software you actually own. The other is available immediately: extract a little more from what has already been sold.

The seat heating was route two. The screen is route two. Both follow the same logic. The hardware sits in the customer's garage, the customer paid for it, and the manufacturer keeps the key.

Cost cuts the customer cannot see are accounting. Cost cuts the customer can feel are a statement about the product.

That is the point where cost reduction starts costing brand. Not the 8,000 jobs. Those are sad, but they are invisible. The 19 seconds are the problem. Because the 19 seconds explain to the customer what they have become: surface area.

And the timing makes it worse. At a 9.8 percent margin, a brand partnership sounds like play. At 2.3 percent, it sounds like desperation. Same action, different balance sheet, entirely different meaning.

The real finding: success was administered

The new chairman of the board of BMW AG is Milan Nedeljković. He took office on 14 May 2026, succeeding Oliver Zipse, who left after 35 years with the company. Nedeljković himself joined BMW as a trainee in 1993.

This is not a criticism of a person. It is a description of a system. A company that missed a structural break is now led by someone who has been inside it for 33 years and previously ran production. Precisely the function where excellence means making known processes better.

For thirty years that was the right competence. From the mid-nineties to roughly 2020, the German premium manufacturer had a recipe that worked: build combustion cars that drive better than other combustion cars, sell them in China where demand grew every year, and steer the margin inside a target corridor.

During that period, success was not a hypothesis to be tested. It was a state to be administered. That shapes organisations more deeply than any strategy. A company that is right for thirty years builds systems that treat deviation as risk. Careers are made through reliability, not dissent. A proposal that questions the recipe has to argue against thirty years of evidence.

Untouchability is not a character flaw. It is a learning effect. A company that wins for a long time learns that its assumptions hold. That is rational. It is also exactly the precondition for not seeing the next break.

Which is how you end up with a product nobody wants. Not because it is badly built. BMW still builds excellent cars by the criteria of 2015. But the criteria changed, and nobody inside had the standing to force the issue while the numbers still looked fine.

By the time the numbers stopped looking fine, it was too late for products. So what remained was the balance sheet.

Not an isolated case

If the diagnosis holds, it should show up next door too. It does.

Mercedes-Benz delivered 210,200 vehicles in China in the first half of 2026, down 28 percent year over year. Audi reached 232,227 vehicles in China including Hong Kong, a decline of around 19 percent. Three brands, same direction, same half-year. That is not a management error at one site. That is a recipe that stopped working for all three at once.

Volkswagen provides the most honest evidence. Cariad, the in-house software unit with around 6,000 people from every group brand, was demoted to coordinator. Software for future models in Europe and the US comes from Rivian. The driver assistance system for China comes from Xpeng starting in 2026. Before that, software problems had delayed models like the electric Macan by years.

That is a complete surrender in the discipline that decides the next decade. A group that masters gearboxes, engines and chassis in-house is buying the operating system of its products from an American startup and a Chinese competitor.

You can call that pragmatic. You can also read it for what it is: six years of trying to build the new thing with the methods of the old one.

Who fills the gap

Gaps do not stay empty. They get taken, by companies that have no success to administer.

BYD registered 26,252 vehicles in Germany in the first half of 2026, up 315 percent, displacing MG as the leading Chinese brand. Chinese manufacturers reached roughly 3.1 percent market share in Germany in the first quarter of 2026, after 2.4 percent for full-year 2025 and 1.7 percent in 2024.

Xiaomi is the real proof. A company that was recently building smartphones ranks third among China's best-selling BEV models in the first quarter of 2026 with the YU7, at 71,623 units. The European launch is announced for 2027.

Nio and Li Auto serve exactly the segment where German brands were unchallenged for thirty years: premium, status, technology as the argument.

The difference is not price. The difference is the clock. These companies have no thirty years of evidence that their existing recipe works. They only have a market that tells them every month whether they guessed right. Their disadvantage of having no history is their one genuine advantage.

What follows for other companies

This is not a car story. Cars are simply the case where you can read it off quarterly numbers. The pattern shows up in every company that was successful for a long time, and it always has the same four symptoms.

One: the margin falls before revenue falls. Revenue is a lagging indicator, because existing customers are slow. Margin is the leading one, because it shows what the market still pays voluntarily for the product. Watch revenue and you notice two years late.

Two: new revenue comes from the installed base, not from the product. If the next revenue step comes from a fee for something that used to be included, that is a diagnostic sign. Not because the fee is wrong, but because it says: right now we have nothing new that anyone pays more for voluntarily.

Three: promotions reward reliability, not deviation. Look at who was promoted in the last five years. If none of them ever publicly questioned a core assumption of the house, the question is not whether the company has such people. The question is where they went.

Four: the word "cost programme" replaces the word "product". In healthy years, communication describes what is being built. In the other years, it describes what is being removed. Compare the press releases from 2021 with those from 2026 and you can watch the switch happen inside the same vocabulary.

The practical question is not: are we saving enough? It is: when did we last sell something that did not exist before?

If you have no answer from the last 24 months, you are administering.

The 19 seconds

It would be too easy to write off the Spider-Man animation as a PR mistake. A mistake would mean somebody failed to think it through. But somebody did think it through. Somebody calculated what that surface is worth and concluded it is worth more than the promise from late 2023.

That is a legitimate business decision. At a 2.3 percent margin it is even understandable.

It is also a confession. Selling your customers' attention says you currently have nothing left to sell inside the product.

And running ads at a purchase price of 100,000 euros forgets what the customer actually expected for that money. Not horsepower. Respect.


Sources

The incident

Numbers and guidance

Cost programme and leadership

Competition

Related Articles

Jonas Höttler builds software for mid-sized companies with Balane.