How many have experienced cpu failure--of any kind? Of those that have, how many could be attributed to some form of radiation degeneration of the cpu (or memory)? Much harder to access as without extreme analysis, there is simply no way to know.
I have experienced failure which, after extreme analysis, we could only attribute to radiation degeneration of a byte of memory - see Re: What do you know, and how do you know that you know it? for some details.
Later in the same job I was writing interrupt handlers for a new ARM-based chipset, and found a problem in that chipset - a bug in certain (characterisable) situations when a multiple-register store instruction crossed a 64k page boundary. Somehow that did less to shake my confidence in the perfection of hardware than the memory corruption problem.