The Closed Loop

I. The Healthy Twin Buckles

When we dissected the Volkswagen board's self-diagnosis, the German car industry still had one counter-example: BMW. The company posting record profits while Wolfsburg wrestled with itself. The living proof that it could be done differently.

That proof expired this week. On 16 June 2026, BMW cut its guidance. Where it had spoken of a „moderate" decline in pre-tax earnings, it now speaks of a „marked" one; and for the first time the board expects deliveries to fall rather than hold steady. Sales in China were down 17.6 percent over the first five months of the year. The investment bank Jefferies halved its expected 2026 automotive margin from 5.2 to 2 percent. The share stands about 31 percent below where it began the year. The 2025 financial year had already brought a 6.3 percent drop in revenue, to 133.5 billion euros, and an 11.5 percent collapse in operating earnings.

But the remarkable thing is not the profit warning. Quarters swing, markets turn; that is the ordinary breathing of an industrial group. The remarkable thing is what runs on, undisturbed, beneath the bad news. For as the substance thins, three movements keep turning that together form a closed loop. The weak quarter is not the story. The loop is.

II. First Movement: The Substance Migrates

BMW is not building its future in Bavaria. China has long been the group's largest research and development presence outside Germany; within three years its forces there were tripled, and more than 3,200 engineers, designers and software specialists now work at four sites in Beijing, Shanghai, Shenyang and Nanjing.

This is no longer the adaptation of a finished car to a foreign taste. It is foundational competence. In China, BMW has built the capacity for the full development of intelligent, connected vehicles — from the chip to the user interface. China is the only market outside Germany where the company runs „Skylab", its human-machine-interaction design team, complete with its own usability lab holding all core design functions. The Chinese team was decisively involved in the conception, testing and validation of the all-electric i5; it is helping to develop the „Neue Klasse", whose local production in China begins in 2026, embedded in a complete ecosystem of research, supply chain and battery manufacturing.

BMW names the driver itself, without the slightest embarrassment: the Chinese government, it says, is „a major supporting factor". That is the true sentence behind the strategy. What is being relocated here is not a plant but the very ability to invent the next car. And it migrates to where a state will carry it.

III. Second Movement: The Taxpayer Provides the Push

One might suppose Germany leaves its corporations to fend for themselves. The opposite is true. While the developmental power drifts east, the public purse at home underwrites the most expensive, riskiest foundations.

In 2020 BMW received 60 million euros from the federal government and Bavaria for battery-cell research — 70 percent federal, 30 percent state — within the European flagship scheme IPCEI, to which the German taxpayer alone contributed 1.25 billion euros. In 2025 a further 273 million euros followed for its hydrogen project, 191 million from the federal government and 82 from Bavaria. The mechanism behind it is the real lever: classification as a „project of common European interest" lets the state subsidise on a scale the normal aid rules would forbid.

Honesty is owed here, and we hold to it: these funds are formally tied to development and first industrial use within Europe. No German tax money flows directly to Shenyang. The point is subtler and weighs more. The public assumes the foundational cost of a competence — cell chemistry, solid-state battery, fuel cell, together with the quiet permanent subsidy of Fraunhofer institutes, universities and research programmes — the very competence the company then exploits globally, increasingly in China. The risk of the foundation is socialised. The return on the application is privatised.

IV. Third Movement: The Profit Flows Upward

And that return has a very precise address. Stefan Quandt holds 27.7 percent, his sister Susanne Klatten 22.5 percent — together nearly half the company. Since 2001 their share of BMW's profits adds up to 58 billion euros, of which 18 billion were paid out as dividends.

The origins of this fortune are not our subject; others have examined them sufficiently already. What concerns us is the present, and it is plain enough. In the very year the company cuts its guidance, it pays a 2025 dividend of 4.40 euros per ordinary share, 2.42 billion euros in all; by their stakes, a good billion of that falls to the two siblings. Earnings fall, deliveries fall, the payout to the owning family stays in the billions.

Where the money goes from there is hard to trace. It lands in holding companies that, because they employ so few people, need publish only an abbreviated balance sheet; one of them was briefly visible in the register and then vanished again. Since 1996 this stream has been taxed at a fraction of the former rate. The company is listed and transparent; the wealth it feeds is not.

V. The Loop

Now the figure can be traced in full. The public bears the push of foundational research. The substance — the ability to develop the next vehicle — migrates to where another state supports it. The profit concentrates in two family holdings whose inner workings escape the public. And the risk, should the bet fail, is borne in the end, as always, by the workforce and the location.

This is no accusation against two individuals. Quandt and Klatten do what any owner in this system rationally does. It is the description of a loop that works without ill intent and draws its force precisely from that. Elsewhere we have called it the husbanding of the many for the good of the few. BMW displays it in a rare, pure form: not despite its success, but at the moment its success crumbles and the loop turns on regardless, as though nothing had happened.

The healthy twin has not fallen ill. It was never the counter-example. It was the same machine — only quieter.