September 14, 2026 | Cars Are Becoming Software Platforms — What to Know Now About the Next Auto Race

For more than a century, the automobile industry competed around physical engineering.

Engines.

Transmissions.

Suspension.

Fuel efficiency.

Manufacturing quality.

Those things still matter.

But the industry's next major competitive battle is increasingly about something very different:

software.

On September 13, Hyundai Motor Group said it will launch vehicles using its own advanced driver-assistance software in late 2029.

That is two years later than the company's original target.

The delay matters.

Hyundai and Kia together sell more than seven million vehicles each year and rank among the world's largest automotive groups.

Yet even a company with that manufacturing scale is finding it difficult to build sophisticated autonomous-driving software entirely by itself.

Hyundai has therefore decided to work more closely with Nvidia.

The companies plan to introduce Nvidia-powered Level 2+ and Level 2++ driver-assistance systems in Hyundai vehicles from 2028.

Meanwhile, Hyundai will continue developing its own software platform, called Atria.

The strategy offers a useful picture of where the global car industry is going.

Automakers increasingly need to be good at two very different things:

building excellent physical machines

and

operating sophisticated software platforms.

That combination may determine which car companies remain competitive in the 2030s.

WHAT HAPPENED

Hyundai originally planned to introduce its proprietary advanced driver-assistance platform around late 2027.

The target has now moved to late 2029.

Instead of waiting until its internal system is ready, Hyundai will use Nvidia technology as an intermediate step.

From 2028, the company plans vehicles using Nvidia's Hyperion 10 platform.

The systems will offer Level 2+ and Level 2++ driving assistance.

These terms generally describe systems that can control more parts of driving than basic cruise control.

They can assist with highway driving and, in more advanced versions, more complex urban situations.

But there is an important limitation:

the human driver remains responsible.

These are not fully autonomous cars.

WHY THE DELAY MATTERS

Traditional automakers are extremely good at complicated physical engineering.

A modern car contains tens of thousands of components.

It must operate reliably for years.

It must survive heat, cold, vibration and accidents.

Building cars at global scale is one of the most difficult manufacturing tasks in the world.

Software creates a different challenge.

Advanced driving systems require:

large quantities of real-world data,

AI models,

powerful processors,

continuous software updates,

simulation,

mapping,

and rapid development cycles.

The culture of software development is also different from traditional vehicle engineering.

Car development has historically operated on multi-year cycles.

Software can change every week.

TESLA CHANGED THE EXPECTATION

Tesla helped popularize the idea that a vehicle can continue improving after it has been sold.

Software updates can add:

new interface features,

driving functions,

battery improvements,

and entertainment services.

That changes the business model.

A car is no longer simply a finished product delivered to a customer.

It can become a platform that continues evolving.

CHINESE AUTOMAKERS ARE PUSHING THE SAME DIRECTION

Chinese EV manufacturers are also competing aggressively in software.

They are integrating:

AI assistants,

smart cockpits,

driver assistance,

navigation,

and connected services.

Product cycles are fast.

Software updates are frequent.

This creates pressure on traditional Japanese, Korean, European and American manufacturers.

HYUNDAI'S RESPONSE IS A HYBRID STRATEGY

Hyundai is not abandoning internal development.

That point matters.

Its strategy is essentially:

use Nvidia now,

learn from the deployed system,

collect data,

then use that experience to improve Hyundai's own software.

Hyundai executive Park Min-woo said the partnership is not about handing the company's future entirely to Nvidia.

Instead, Hyundai intends to co-design technology and use vehicle data to improve Atria.

WHY DATA MATTERS SO MUCH

Autonomous-driving AI improves by seeing more driving situations.

A vehicle may encounter:

pedestrians,

construction zones,

rain,

snow,

unusual road markings,

motorcycles,

emergency vehicles,

and millions of other edge cases.

No engineering team can manually design rules for every possible situation.

AI systems learn partly from data.

That gives companies with large vehicle fleets a potential advantage.

HYUNDAI AND KIA HAVE SCALE

Together, Hyundai and Kia sell more than seven million vehicles annually.

The group wants to use that global fleet to collect driving data.

Hyundai says it hopes to surpass competitors in accumulated driving data by 2033.

Whether it achieves that goal remains uncertain.

But the strategy highlights an important change:

cars are becoming data-generating devices.

THE AUTOMAKER'S NEW FLYWHEEL

A future vehicle business could work like this:

sell cars,

collect driving data,

use the data to improve AI,

send improved software back to vehicles,

make the vehicles more attractive,

sell more cars,

collect more data.

Technology companies often call this a flywheel.

The more users a platform has, the more data it gains, potentially improving the product and attracting even more users.

THIS ALSO CREATES A PRIVACY QUESTION

Vehicle data can include highly sensitive information.

Location.

Driving habits.

Camera images.

Routes.

Potentially information about passengers.

Countries will need rules governing:

what data can be collected,

how long it can be stored,

where it can be processed,

and whether it can be used to train AI.

WHY NVIDIA IS BECOMING IMPORTANT TO CARS

Nvidia began as a graphics-chip company.

Its processors later became important for artificial intelligence.

Now the company is expanding into vehicles, robotics and industrial systems.

Modern driver-assistance requires large amounts of AI computing.

Cars increasingly need powerful processors capable of analyzing information from:

cameras,

radar,

ultrasonic sensors,

and potentially lidar.

Nvidia wants to provide both the computing hardware and the software platform that connects these systems.

THAT CREATES A NEW DEPENDENCY

Traditional automakers historically controlled much of the core vehicle technology.

If the intelligence of the vehicle increasingly depends on an external chip and software company, automakers risk becoming dependent on technology suppliers.

This creates a strategic question:

How much software should an automaker build internally?

BUILD EVERYTHING YOURSELF?

The advantage is control.

The manufacturer owns the technology.

It can differentiate the product.

It keeps more economic value inside the company.

The disadvantage is cost and speed.

AI talent is expensive.

Software development is difficult.

Competitors may move faster.

BUY EVERYTHING FROM A TECHNOLOGY COMPANY?

This can accelerate development.

But if many automakers use the same technology, their vehicles may become harder to differentiate.

The technology supplier may capture more of the profits.

And changing suppliers later can become expensive.

THE LIKELY ANSWER IS A MIX

Hyundai's strategy may become common.

Automakers will probably own the parts of software that differentiate their brand while buying or partnering for infrastructure that would be expensive to recreate.

This already happens in other industries.

Smartphone companies do not manufacture every chip or network component themselves.

Cloud companies use components from many suppliers.

Cars may follow a similar model.

SENSORS ARE ANOTHER COMPETITIVE CHOICE

Hyundai's Nvidia-powered vehicles will initially avoid expensive lidar.

Instead they will use:

cameras,

radar,

and ultrasonic sensors.

Lidar uses laser pulses to create detailed three-dimensional maps of the environment.

It can improve perception but adds cost.

Tesla has famously relied heavily on cameras.

Other companies use lidar.

There is still no complete industry consensus on the best sensor combination.

SAFETY CHANGES THE ECONOMICS

Smartphone software can crash and be annoying.

Vehicle software can affect physical safety.

That means automotive AI needs unusually high reliability.

Systems require redundancy.

If one sensor fails, another may need to continue providing enough information for safe operation.

This is one reason autonomous driving has developed more slowly than early forecasts suggested.

THE LEVEL NUMBERS MATTER

Consumers should be careful with marketing terms.

Level 2 systems can perform steering and acceleration under certain conditions.

The driver must still supervise.

Level 3 systems can take over more responsibility under defined circumstances.

Higher levels move toward full automation.

A car that can drive itself on a highway does not necessarily have the ability to drive itself everywhere.

WHY THIS MATTERS FOR JAPAN

Japan has one of the world's largest automotive industries.

Toyota, Honda, Nissan, Subaru, Mazda and many suppliers employ large numbers of people.

The industry's traditional strengths include:

manufacturing quality,

reliability,

materials,

mechanical engineering,

and supply-chain management.

Those strengths remain valuable.

But competitive advantage increasingly also requires:

software engineers,

AI researchers,

data infrastructure,

cybersecurity,

and semiconductor expertise.

THE JOB MIX WILL CHANGE

The automobile industry will still need mechanical engineers.

But it will need more people working in:

software,

AI,

data,

chips,

cybersecurity,

sensors,

and cloud systems.

Technicians will also need new skills as cars become increasingly electronic and connected.

This does not mean every traditional automotive job disappears.

It means the industry's center of gravity changes.

A NEW REVENUE MODEL COULD EMERGE

Automakers traditionally earn most revenue when the vehicle is sold.

Software creates possibilities for recurring revenue.

A customer might later pay for:

advanced driving features,

navigation services,

entertainment,

fleet management,

or other connected functions.

This is attractive to automakers because recurring revenue can continue for years after the car sale.

BUT CONSUMERS MAY RESIST

People may not want a car filled with subscriptions.

The industry will need to find a balance.

Charging monthly fees for genuinely new services may be accepted.

Charging recurring fees for hardware already installed in the vehicle may create backlash.

CYBERSECURITY BECOMES CRITICAL

A connected vehicle is also a computer connected to the internet.

That creates cybersecurity risk.

Manufacturers will need to protect cars from:

unauthorized access,

malicious software,

data theft,

and manipulation.

Automotive cybersecurity will become a major field in its own right.

WHAT MAY HAPPEN NEXT

Scenario One: Technology suppliers become extremely powerful.

Nvidia and other platform providers supply core AI systems to many automakers.

Car companies focus increasingly on design, manufacturing and brand.

Software platforms capture a larger share of vehicle economics.

Scenario Two: Major automakers successfully build their own platforms.

Companies such as Hyundai, Toyota and others use partners temporarily but eventually control their own core software.

The automobile industry retains more technological independence.

Scenario Three: The market splits.

A small number of companies build full internal stacks.

Most others rely heavily on technology suppliers.

This may be the most realistic outcome.

Scenario Four: Autonomous driving develops more slowly.

Safety regulation, technical limitations and consumer concerns keep fully autonomous vehicles limited to specific locations and use cases.

Driver-assistance improves steadily without eliminating drivers entirely.

WHAT IT MEANS GLOBALLY

The auto industry is moving from a mechanical competition toward a combined mechanical-software competition.

China has strong EV manufacturing.

The United States has powerful AI and software companies.

Japan, Germany and South Korea have deep manufacturing expertise.

The winners may be the companies and countries that combine both worlds rather than dominate only one.

WHAT IT MEANS FOR BUSINESSES

Automotive suppliers should ask whether their products remain important in a software-defined vehicle.

Some traditional components may become less important.

Others become more valuable.

Demand could grow for:

sensors,

power electronics,

chips,

cooling systems,

network components,

and cybersecurity.

WHAT IT MEANS FOR WORKERS

You do not necessarily need to leave the automotive industry to benefit from AI.

A strong strategy may be to combine existing knowledge with new digital skills.

Examples include:

mechanical engineering + data,

vehicle maintenance + electronics,

manufacturing + AI,

quality control + computer vision,

and logistics + automation.

WHAT READERS CAN DO NOW

When evaluating a new vehicle, do not only compare horsepower and battery range.

Also ask:

How long will software updates be provided?

Which driver-assistance features require supervision?

How is vehicle data handled?

Which features require subscriptions?

What happens if the manufacturer stops supporting the software?

These questions will become increasingly important as vehicles remain on the road for ten years or more.

WHAT YOU CAN DO NOW

The simplest way to understand the next automotive era is to stop thinking of a car as only a machine.

Think of it as three products combined:

a physical vehicle,

a computer,

and a continuously updated service.

When you read automotive news, watch all three layers.

For workers, identify one digital skill that complements your existing physical or industry expertise.

For businesses, examine which parts of your value chain become more important when vehicles rely more heavily on software.

And for consumers, remember that advanced driver assistance is still assistance.

Check the actual capabilities and supervision requirements rather than relying only on product names.

Hyundai's decision to use Nvidia while continuing to develop Atria is not simply a delay in one company's autonomous-driving project.

It captures the central challenge facing the global car industry.

Building excellent cars is no longer enough.

Automakers must also become excellent software companies — without forgetting that their software operates a machine carrying real people on real roads.