Either am I. 
And I get what you’re saying.
But my approach to derating and thoughts on reliability are just the way I have always designed my products. By taking all failure mechanisms into account and designing conservatively I can prevent issues that might occur some time in the future. Especially in a case like this where the operating environment is uncontrolled and hostile.
While it does seem that 63V might be high enough for most 12S ESC use I hesitate to say it will be high enough to be ensure reliable operation of those caps in all setups.
I’m also not sure we can say that no ESC failure has been caused by damaged caps. I don’t know a lot about cap aging and damage mechanisms but they can be somewhere between 100% okay or completely blown with a popped top.
I believe there are scenarios where they are less effective or could even have blown an internal conductive path without popping their top. I know they can go very leaky (high self-discharge) and there might be damage that can be capacitance-reducing too, allowing voltage transients to rise to higher levels.
Something I would think about is one of the caps having a leg snap due to repeated bending or vibration. We now have a lot less capacitance taming the inductive spikes and the voltage rises. Do we have enough voltage rating headroom for that? Is that even a failure we want to survive (by paying for higher rated caps)? If yes, how? More caps? Higher voltage ratings? Just thoughts I have when designing.
In some cases I might just choose to use 63V rated caps. It depends on the client, the MSRP of the product, the marketing, the reputation of the company, and the size/weight/cost of the different cap options. I would sure be wanting to use the 75V cap though….with an appropriately low ESR, appropriately tight tolerance, appropriate temp rating…blah, blah, blah.
If we feel that something shouldn’t happen, that something shouldn’t fail, without a boatload of supporting testing then IMO we are setting ourselves up for a product reliability disaster. Maybe not now. Maybe not for quite a while. But it will happen at some point.
For my clients and I this is unacceptable. But I completely respect anyone’s decision to use components at closer to their ratings than I would. There’s no “correct” cap voltage rating, operating temperature rating, or vibration/shock tolerance/limit for caps used in an ESC. There’s just our personal design methodology/approach and any limits on cost, size, etc., that we choose to follow.
For me, I like to not worry as much about battery lead length and how close the leads are kept together (to minimize inductance) as others might have to. It leaves more time to worry about heating, vibration, etc. 
