My recollection is that 5/100K was an action level the epidemiologists were working with early on - that certainly has probably changed given that the bug has ended up more infectious than originally estimated. But OTOH, the original concept here being that if you can contact trace all the positives, you can keep it in check, and so the stable infection level really has more to do with how much public health resource you have to do contract tracing rather than any particular property of the infection and that is what I believe that 5/100k number was based on.
Now in practice Covid has overrun anyone's capability to do effective contact tracing so that has not been part of the conversation for months, but that doesn't mean the concept wasn't correct. As this comes under control there should be levels at which things like contact tracing (or whatever) that were not feasible at high infection rates come back into play, and those considerations should allow the Epids to come up with action levels to relax. But the larger point is that it's not all just black magic and numbers pulled from a hat. The estimates may not always turn out to be correct but they do have non-arbitrary reasons for choosing them.