Relatively speaking, I would bet that most of the 6700 XTs wouldn't last as long as the other RX 6000 GPUs since the 6700XT is clocked significantly higher. By degrade, I would expect at least some of the 6700 XTs that make it into gamers hands to no longer conform to spec in about 6 months to a year of normal gaming loads every day, maybe 4 or more hours a day, basically not "game for 1 day, rest for 6 days", since I know there are those people as well. I do expect there will be significantly more RMA rates for 6700 XTs versus the RX 6800.
This goes for almost anything, CPUs, RAM, etc. In a given generation/process node, there are yields. As an example, I will talk about RX 6000 GPUs as a whole. According to TechPowerUp's GPU database, all RX 6000 GPUs are running TSMC's 7nm node, and 6800+ is using Navi 21, 6700 XT is using Navi 22. But the transistors are the same, just, AMD isn't putting as many in Navi 22. The higher you clock a GPU, the lower yields you will get, that is just how engineering works. This process is called binning, and the GPU processors that can handle the higher speeds won't fail as quickly because while they all degrade at nearly the same rate (in the same generation/process node) there is a spec, and anything that meets spec or exceeds spec will be fine. However, if you spec a GPU w/ a significantly higher clock speed, you increase the silicon quality requirement needed to achieve that, and here is where it can get tricky. It either goes 1 way or the other: either the GPU with lower clock will be super stable and have almost zero failure rate and the GPU with higher clock is stable enough, OR the GPU with lower clock is stable enough, and the GPU with higher clock will have noticeable failure rates. The higher you clock a product, it will be less stable (given the same voltage). Sure, you can go from 2200 MHz to 2300 MHz on the same voltage, it just means you didn't actually need as high a voltage for 2200 MHz. The higher the clock speed, usually more voltage/current, and in the end, you want more cores, not more clock speed. For example, let's say AMD made a Navi GPU with 5120 cores and 1500 MHz core clock speed, that would be many times more reliable than making a Navi GPU with 2560 cores and 3000 MHz core clock speed, because after a certain point, there is diminishing returns with overclocking. There are more factors into play obviously, like cooling, current draw, etc, but all I was saying was, I expect the 6700 XT to have higher failure rates than the rest of the RX 6000 GPUs.
Also, 6800 XT and 6900 XT don't have as high base clocks; 6800XT and 6900XT GPUs are around 1800-1900 MHz base clock, not game clock.
Smear campaign? Um... where are you getting this bullshit from?
If yields are better on Navi 22, then alrighty, it won't degrade as quickly as I thought it would. I thought it was running on the same Navi 21 GPU. But yeah, generally speaking, higher clock speeds = higher voltage, and that's why I was saying it would degrade quicker, because of the higher voltage.
