This is only partially true. Yes we do engineer things to fail at a certain point, but that’s only because back in the day we naively assumed that we could engineer things not to fail at all.
Yes a stator of an electric engine will probably not fail for 100 years, but the seals will - yes the statically stressed metal part will hold until it crumbles to rust, but the dynamically stressed plastic part won’t - yes the silicon in an IC-Chip is protected from corrosion, but the connector pins aren’t.
The point I’m trying to make is that there’s always a part that will fail before another, there’s no way to economicaly engineer around that, today we simply have the data to statistically define a failure point.
A fridge usually has a 10 year warranty. This isn’t even the end of life point. After 10 years it’s most likely that 80-90% of devices will still work. This means that if your device survived 10 years it will most likely work for another 5-10 years.