You paid.
But do you really own it?

It used to be fairly simple.

You bought software on a floppy disk or CD, installed it on your computer and — as long as the hardware still worked — could use it even if its maker disappeared, was acquired or simply lost interest in the product.

Games worked much the same way. A disc bought twenty years ago did not need permission from the publisher’s server before you could play.

Today, we increasingly buy something that looks like a product, costs like a product and is advertised as a product, but technically behaves more like a service rented for an unspecified period.

And this is not even about conventional SaaS.

SaaS itself is not the problem

Software as a Service makes a great deal of sense.

Microsoft 365, accounting systems, CRMs and cloud tools are sold as services from the outset.

I pay a subscription. I get access. I stop paying — I lose access.

I may like that model or not, but the rules are relatively clear.

The problem starts elsewhere.

When I pay once for a product that, in ordinary terms, I have bought, but its maker retains the ability to disable its core functionality by closing a server.

That is when the boundary between ownership and services gets interesting.

The game you bought that no longer exists

A good example is The Crew from Ubisoft.

The game was released in 2014 and sold as an ordinary product. You could buy it for PC and consoles.

On 14 December 2023, Ubisoft announced it was withdrawing the game from sale. Existing owners could continue playing until 31 March 2024.

Then the servers were switched off.

That was it.

Not “no more multiplayer”.

Not “no more updates”.

The game became inaccessible on every platform. Ubisoft cited changes to server infrastructure and licensing constraints.

The publisher did not stop providing an add-on. It stopped people using the game itself.

“Buy” does not always mean “own”

Films and television face a similar problem.

In 2023, Sony told PlayStation users that licensing changes meant some previously purchased Discovery content would be removed from their libraries.

Not rented. Not part of a subscription. Purchased.

In the end, the removal did not happen: Sony later announced that revised licensing terms would keep the content available.

Good.

But the fact that this scenario was possible at all exposes the problem.

If I “buy” a film, but continuing to watch it depends on a future agreement between two companies that I cannot influence, what exactly have I bought?

• A file?
• A licence?
• An access right?
• A promise that I will probably still be able to watch it in a few years?

The shop interface usually just says: BUY.

Now take it a step further. A car.

With a game costing PLN 200 or 300, you might still shrug your shoulders.

It is harder when similar dependencies appear in objects costing tens of thousands of dollars.

In 2024, the American electric-car manufacturer Fisker.

The company had launched the Fisker Ocean SUV, but financial, organisational, servicing and software problems led to Chapter 11 proceedings and then liquidation. The Ocean itself also suffered numerous software problems from the beginning.

That raised a question almost nobody had needed to consider when buying a car fifteen years earlier:

What happens to the cars when their manufacturer’s IT infrastructure disappears?

During Fisker’s liquidation, American Lease was buying more than 3,000 Oceans.

At one point, the deal was in doubt because the data and services needed to support the vehicles were tied to Fisker’s cloud infrastructure, and moving them to new servers was not straightforward. The issue was serious enough to require an additional agreement on continuing technical support.

There is the app. The backend. The user account. Cloud services. OTA updates. Diagnostics. Certificates. Data connecting a car to its owner. Fleet-management systems.

Suddenly, a manufacturer’s bankruptcy no longer means only:

“It will be harder to buy an original headlight.”

It can mean:

“Some of the infrastructure needed to support my physical possession has just lost its owner.”

Fisker owners had to become part of the manufacturer themselves

The story has a particularly interesting sequel.

Ocean owners organised themselves into the Fisker Owners Association.

The community began taking on tasks a manufacturer would normally handle: sourcing parts, organising servicing, developing software tools and alternative systems to keep the cars alive.

By 2025, the organisation had more than 4,000 owners and was developing its own ways to access some vehicle functions, among other things.

It is an extraordinary example of community engineering.

It is also a rather absurd situation.

You bought a car. The manufacturer collapsed. So you and several thousand other customers organise something like a substitute manufacturer to keep the car properly supported in software.

A manufacturer failing did not used to mean its product failing

This may be the biggest change.

Imagine a car built in 1995.

The company might go bankrupt. The dealer might disappear. Parts might become harder to find.

But if the car is in your garage and works, you do not need its manufacturer every morning to start the engine.

The same was true of most software.

You can find an old computer, install a program from its original media and run it without its author’s involvement.

Today, more and more products have an external dependency the user does not control.

And it need not be a conspiracy by the maker.

Servers cost money to run. Licences expire. Security standards change. Libraries lose support. Companies go bankrupt. Services become unprofitable.

These are all real problems.

But from the customer’s perspective, the result is the same:

A product they paid for can stop working for reasons entirely unrelated to the condition of that product.

Stop Killing Games

The shutdown of The Crew helped trigger the Stop Killing Games.

The idea is not simply that “the company must run the servers for another hundred years”.

That would be economically absurd.

It is more about this: when a manufacturer ends support, it should, within reason, leave the product in a state that allows continued use.

That might mean an offline mode, the ability to run your own server, peer-to-peer connections, releasing the necessary tools, or another technical solution anticipated when the product was designed.

The issue has moved far beyond an online petition. The European Citizens’ Initiative “Stop Destroying Videogames” collected 1,294,188 verified statements of support and was formally submitted to the European Commission in January 2026.

The Commission published its response on 16 June 2026.

For now, it has not proposed a legal obligation to keep games playable after their commercial life ends. It has, however, announced discussions with industry and consumer organisations on an end-of-life code of practice and further analysis of existing consumer rights.

Interestingly, the Commission’s own analysis recognises much the same problem: shutting down servers can make a product bought for a one-off payment partly or entirely unusable. It also identifies solutions such as offline versions, private servers and P2P connections.

The discussion is clearly no longer just a matter for a few angry gamers.

A service should be a service. A product should be a product.

I have no problem with Netflix.

I pay monthly for catalogue access and know a particular film might disappear from it.

I have no problem with a VPS.

I stop paying — the server disappears.

I have no problem with an online service whose very purpose depends on the provider’s running infrastructure.

The trouble starts when the sales language says “you are buying a product”, while the technical architecture says “you are renting access for as long as we let you connect”.

Perhaps the answer is not to ban everything that uses the cloud.

That would be completely unrealistic anyway.

Perhaps the answer is much simpler:

Design the end of a product’s life at the beginning of it.

If a game needs a server, perhaps it should have a plan for working after that server closes.

If a physical device needs the cloud, perhaps its basic functions should remain available if the manufacturer disappears.

If purchased digital content is really a renewable licence dependent on third-party agreements, perhaps “Buy” is not the best label for what we are getting.

Increasingly, we are no longer just buying things

We are buying an entire chain:

One missing link is enough.

Technically, it is fascinating.

As a developer, I can appreciate what network infrastructure, OTA updates and service architectures make possible.

As a user, though, I would like to know one thing:

Will the thing I have just bought still be mine when the company that sold it ceases to exist?

Perhaps the biggest change of the past fifteen years is not that everything has become a service.

It is that it has become harder to notice the moment when we stopped really owning something.