The car is becoming a supercomputer on wheels

The most powerful computer you own may soon be parked outside your home.

For most of automotive history, a car was a mechanical marvel with a few electronics added on. Then came the electronic control unit. One box for the engine. One for the brakes. One for the airbags, the windows, the seats, the radio. Each added by a different supplier, each with its own software, each solving its own small problem.

By the time software-defined vehicles became a boardroom phrase, a premium car could carry around a hundred of these boxes and, by industry estimates, some 100 million lines of code. A network of small computers, stitched together by kilometres of wiring.

That model is ending. The car is becoming a supercomputer on wheels.

From a hundred boxes to a few brains

The shift is simple to describe and hard to execute. Instead of dozens of separate control units, the car runs on a handful of high-performance computers, connected to "zone" controllers that handle everything in one area of the vehicle.

Think of it as the journey from a drawer full of gadgets to a smartphone. One central brain, many apps, and updates that arrive over the air.

Several forces push in the same direction.

• Features. Driver assistance, automated parking and smarter cockpits need far more computing power than small control units can provide.
• Speed. Updating one central platform is much faster than coordinating dozens of suppliers for every change.
• Cost and weight. Fewer boxes and simpler wiring mean lighter, cheaper and easier-to-build cars.
• Electrification. Electric platforms are designed from scratch, which makes it easier to start with a new architecture.

In my earlier essay on spaltmaß and software, I wrote that you can often read a company's organisational structure in its user interface. The same is true of its electronics. A hundred boxes reflect a hundred separate decisions. A central computer reflects a company that has learnt to think as one.

China changed the clock

The pace of this shift is now being set in China. Advanced driver assistance that can navigate city streets has moved from a premium option to a mainstream feature in a remarkably short time. Chinese carmakers are shipping new software in months, not years, and building their own chips, operating systems and assistance stacks.

That speed is not only about technology. It is about organisation, local ecosystems and a willingness to own the whole stack. For global carmakers and suppliers, it has turned a technology roadmap into a competitive deadline.

Who owns the brain?

When the car gets a brain, the obvious question is who builds it. The contest is open, and the field is crowded.

• Chipmakers bring the silicon and increasingly the software around it, from driver assistance platforms to reference operating systems.
• Traditional suppliers bring decades of safety engineering and deep knowledge of how cars actually behave on the road.
• Carmakers want to own the software that defines the driving experience, and have built large in-house teams to do it.
• Technology companies, especially in China, offer the full stack, from chips to cockpit to cloud.

Each has a credible claim. None has the whole answer. The likely outcome is not a single winner, but a new kind of coopetition in which carmakers decide which layers to own, which to buy and which to share.

The seam between the brain and the body

There is a quieter battle underneath. A central computer can decide what the car should do: brake here, turn there. But brakes, steering, motors and suspension still have to execute that decision, safely, in milliseconds.

The software layer that translates intent into coordinated movement, deciding which actuator does what when several systems ask for different things at once, has become one of the most contested places in the vehicle. It sits between the brain and the body. Whoever owns it shapes how a car feels to drive, and captures some of the rare software margins left in a hardware business.

In a software-defined vehicle, value does not disappear from hardware. It moves to the interfaces.

The first mass-produced robot

Step back and the picture becomes clearer. A modern car senses its surroundings, plans a path, and moves its body through the physical world. That is the definition of a robot. Cars are simply the first robots we have learnt to build by the millions.

Which is why the lessons matter far beyond automotive. Humanoid robots face the same questions. A central brain or many small controllers? Who owns the software that turns intent into motion? Which players control the chips, the actuators and the data? The industry that works out the software-defined vehicle first will have a head start on the machines that come next.

Where to play

For suppliers and carmakers alike, the strategic questions are becoming sharper.

• Will you build the brain, the nervous system or the muscles, and what will you deliberately leave to others?
• Can your software run on someone else's hardware, and are you willing to sell it that way?
• How will you earn money over the life of the vehicle, not only at the moment it is sold?
• And can your organisation move at the speed the market now demands?

The car is becoming a computer. The harder transition is for the companies that build it to start thinking like computer companies, without forgetting why people trusted them with their lives in the first place.