I’m not an expert, so take what I say with a grain of salt, but here’s what I feel has been more and more the case… It’s interesting that the examples you set are all hardware communicating with other hardware. That is a key point because any company selling you those devices can easily defend themselves legally if you decide to sue them for using your data just by saying “how else would we get the device working? It is fundamental to read your data to make the device do what is advertised for” and the case would be dropped faster than my dog comes when I open his food. Now imagine the keyboard company is caught sending the key strokes to their servers… Without a good terms of use contract they would lose immediately against legal action. And even a terms of use contract might be considered null if it is proven to be abusive or something.

When it comes to a software company things get a lot blurrier. It’s harder to define the needs for some actions and how things could work vs how they work. So I think it’s not uncommon to have this kind of clauses in such cases, specially for getting user data for maintenance and so on. It was less common in the past but as there are more practical cases and experience of where the law draws lines and limits this kind of additions and edits of user contracts are becoming pretty normal.