A working device stops receiving software updates while it is still perfectly capable of running. The cutoff is a supply-chain and testing decision made well upstream of the user.

Support ends at the chip before the product

The processor and radio components in a device come with vendor support windows of their own, and those windows are agreed when the part is designed.

Once a chip vendor stops issuing updated low-level code for a part, the device maker cannot rebuild a current operating system on top of it.

The product's public support period is therefore bounded by a contract signed years earlier, which is why devices launched late in a chip's life age quickly.

Drivers are the hardest part to carry forward

Operating systems change the interfaces that hardware drivers use. Every change requires the driver to be updated by whoever holds the source, which is often not the device maker.

Where that code is closed and unmaintained, the only path forward is to keep the old interface alive, and that burden grows with each release.

Projects that maintain long-lived community builds run into exactly this wall, which is why they often ship with certain hardware features disabled.

Testing cost scales with the model count

Each supported device needs its own validation run before a release ships, covering the camera, the modem, power management and the specific quirks of that board.

That work does not shrink as a model gets older, but the number of people still using it does, so the cost per user climbs steadily.

Manufacturers with fewer models sustain longer support windows for this reason alone, independent of how generous their policy sounds.

Requirements rise for reasons unrelated to speed

Newer releases often demand specific security hardware, such as a dedicated element for storing keys, rather than simply more processing power.

A device without that component cannot meet the guarantees the new version makes, so it fails the requirement regardless of how fast it runs.

This is the mechanism behind cutoffs that appear arbitrary to owners whose machines feel entirely responsive under the older version.

Losing updates is not the same as breaking

An unsupported device keeps working, and for tasks that do not touch the internet it may serve for years without any practical downside.

The real exposure is in the browser and in anything handling untrusted data, where unpatched flaws become known and reachable over time.

Separating those uses, rather than replacing the hardware, is the honest response to a support window closing on a machine that still functions.