I want you to go to AMDzone and post this somewhere, anywhere, on their board, just to stir them up. That would be hilarious.
Seriously, AMD has historically produced processors that run hot? What? There was the T-bird but other than that . . . are you sure you aren't confusing them with Cyrix?
. . . and you're STILL spreading the FUD from the Tom's Hardware video in which an Athlon XP burned out when the HSF was disabled? Do you have any idea how long it has been since AMD produced processors that can die like that, much less take a board with them in the process?
. . . and your employer is complaining about failure rates among x86-64 capable processors from AMD after 4 years of usage? I have a fairly early processor from that generation (Sempron 2800+, s754 . . . yes I know x86-64 was superficially disabled on that chip, but it was the same uarch as the 130nm Hammers on the then-new 90nm process and minus some L2 cache) that I put into a system in 2005 and IT STILL RUNs. I ran it overclocked to 2.32 ghz for years, using the stock HSF. Unreliable my ass. Athlon 64s, X2s, and the like do not have a rep for high failure rates. Who was the vendor or OEM who supplied you with your systems?
Furthermore, most - if not all - of those AMD processors about which your employer is complaining should be able to underclock themselves. It's called Cool n' Quiet.
The competing processors Intel sold from those days were based on Netburst, and when it came to Prescotts and Smithfields, PLENTY of them either failed or just throttled like hell at stock because OEMs had problems keeping them cool.
FYI, the unlock rate on Phenom II X2s and X3s to quad-core processors has been calculated to be around %72.9, possibly by less-than-scientific methods, but that's the number that's bandied about in circles where people care about such things.