Im not in agreement with this. Coffeelake, alderlake and to a lesser extent arrowlake have been really good generations.
I dont see any masking of progress or super obvious power consumption issues. Their biggest problem is the die size of the pcores, not their power draw. Zen 4 core for core (so 8vs8) is around 8% more efficient than 8 alderlake pcores at iso power in something like cbr. Im fairly confident raptor and arrowlake are more efficient than zen 5 core for core, not to mention that they also ship with a lot more cores.
I think people who think intel is inefficient are just biased or cant see the forest for the trees. In the midrange intel has been smacking everything for the past few years in efficiency thanks to increased core counts.
I've had coffee lake. It was NOT a great gen. Poor IHS contact / delidding ring a bell? And that problem has taken its sweet time to fix too. Until recent generations we've seen major IHS problems.
Alderlake similarly, was NOT a great gen. The only CPUs that did well were doing nothing new, the ones that lacked E cores. Scheduling issues happened a lot and were hard (if not impossible) to fix. The real fix was just disabling half your silicon. I mean lol.
Intel's problem isn't the die size of its P core.
Its the baseline core design. They reworked that. It improved a bit. But they still haven't gotten out of their P/E nonsense hole. Meanwhile a competitor can do
better in every way (scalability, efficiency, linearity in performance, heat/power management / consistent clocking, etc.) with a simplified design that is just built up out of solid parts. A lean core, perhaps some extra cache, and an interconnect.
Intel has none of that in any kind of coherent structure right now. I'm not contesting your statement or the actual performance of a recent Intel part. I'm looking at how they achieve that product to work as it does: its a fucking mess and it doesn't scale, has no future, and is an architectural shitshow. That is why its a complete toss up every gen what's going to break on your Intel CPU this time, and it has, indeed, been several things over the last years. Fundamentally broken things, or things to on the edge of the spec, it shouldn't be happening.
So yeah. A mindshare problem indeed, but it really is caused by the obvious writings on the wall, that have led to actually failing products in real life already. This company is intellectually bankrupt at this point, grasping at straws to meet their yearly target... and still not meeting it.
I think it's main problem is who's going to upgrade to it any enthusiasts that was already on LGA 1700 it isn't worth the platform switch same with those already on AM5 I think most Intel buyers who haven't jumped ship are waiting for Nova Lake keep in mind people care a lot more than they should even about a 5% difference in gaming perfomance it's why even with 14nm++++++++ Intel held on strong all the way through 10th generation. I honestly think Intel hasn't fully recovered from Rocket Lakes reception honestly
Sure there will be some diehards that will buy it regardless, arrow lake reception after they basically canceled meteor lake on desktop meaning the socket will only get one architecture with at best a refresh isn't surprising to me. It's things like that that Intel has fumbled if Meteorlake was awesome and on time lga 1851 would be more well regarded.
Also the i5 which gets people in the door is still expensive and decently worse than the 285k that isnt getting people excited for the platform in the first place.
Which just leaves the decently priced i7 to try and carry the platform.
Just like DLSS and RT performance has become important on the gpu side I believe people at least thinking they are going to get 2-3 generations on a socket is important. I wouldn't invest into lga 1851 for that specific reason even though I don't dislike the cpus and think the 265k is pretty damn solid at it's current price.
That's just my thoughts on it though.
Still happy to see good pricing on the cpu side it's the only thing kinda going for gamers in 2025 like for 1440p gaming 180 ish usd gets you a cpu good enough you'd likely not notice anything above it besides some very niche scenarios like wanting to drive a 480/500hz panel.
Good MT perfomance can be had at less than 300 usd if that floats a person's boat also awesome.
I really think the market loves simplicity and will always lean towards simplicity. That is how conversations work too, right.
People make a topic and they ask 'What's the best gaming CPU'. Nobody is going to come up answering with a concise, nuanced top ten with reasons attached. They'll say 'Get an X3D'.
Similar things applied to Intel: 'get a 3570k'. Get an i7 with HT, whatever it is. Intel does not have those 'get anything like that' products anymore. You have to get specific and really get a perfect CPU for your specific use case AND then you might also want to tweak it a bit if you're in the performance segment with your use case. They created that problem themselves, when they introduced their P/E strategy.
Nobody got time for that shit, they'll save the time and get an X3D instead. It'll also probably save them a bit of power along the way, and the performance they lack in the rare parralelized multithreaded task, is negligible for 95% of the use cases they have for it. That's really all it is.
