My suspicion is a minor bug in the ATmega16U2 bootloader / FLIP. Reprogramming it would add a bit of anecdotal evidence to bolster my suspicion.
Is there any other hardware related, or wiring issue that could have caused the failure?......."Well, 12V applied directly to the ATmega16U2 would certainly cause it some distress."
Maybe the USB/Serial chip on the Arduino is not cut out for 365/24/7 operation?..............."If that were the case, Atmel would have long ago been out of business. They guarantee that each processor will perform within specifications for 20 years (or more).On a more local level, the forum would be flooded with unhappy customers."
Are you suggesting that there may be a way that Vin could make its way to this chip?
For all intents and purposes, that might make someone pause before using them in a situation where they must run 365/24/7. There are a few other threads reporting USB failure, all without definitive resolution.