Yes. If you turn off a computer it can't be compromised.
But I think the interesting question is if its possible have a system that does something useful, while at the same time be 100% secure.
The answer is probably yes, especially small systems that have been designed for a single purpose, but it would be incredibly difficult to prove it. Just because it hasn't been compromised yet doesn't mean it won't in the future.
Even doing a formal proof only proves things within the abstract model the proof works with.
Just because there are other things that are worse, or at least as bad, doesn't mean we shouldn't call put bad behaviour when it happens. And written MS it keeps happening.