So in other words, about as expected performance-wise unless you ate up all the overhype AMD pushed out. Not bad CPUs, but a lot more hype than actual gains compared to the competition and Raptor Lake is right around the corner. That's when it will get actually interesting with the comparisons. I got to say, though, those new temps are ridiculous. I can use a budget air cooler for a freaking 12700k without problems, and a 12900k can do fine with a big air cooler or moderate AIO as well, but even a 7600X would need a 360mm radiator to stay at 90°C according to GN's videos and the 7950X sits at toasty 95°C with their 360mm AIO. That's unacceptable. I really hope Intel will not go overboard in this department, though the 13900K will, unfortunately, most likely pull even here...
I'm a bit surprised, though, that DDR5-6000, while giving quite a nice boost (though obscured by PBO), isn't doing all that much over 5200 for the 7950X. It seems like it actually works better for Intel, considering that the 12900K with OC to 5.1GHz (so all-core boost set to single-core boost frequency) is pretty much equal to the 7950X with PBO despite being the older chip. I guess we need to look at the 7900X for a final judgement here. All in all, people might want to stick to cheaper RAM, though. Total platform cost is also really not great compared to Intel this time around, with boards and RAM being so expensive. All in all, I find my decision to get the 12700K confirmed. AMD fans should rather go with the 5800X3D unless they really need more cores and clocks for productivity.
Also, it would be really great to see some efficiency comparisons for gaming. People still act as if Alder Lake use 100W+ for gaming or some bs like that, when that most definitely not the case. Gaming should be about the most common use case for CPUs in home systems, so it really should be considered in reviews like this when it comes to power consumption. In this regard, other outlets like Igor's did some pretty thorough testing that showed that maximum possible energy draw is a horrible measure for gaming power consumption. I think it would be great to not only have the maximum draw, but also gaming consumption measured. Just as a general suggestion for making the testing more thorough.