> They have decades of X86 experience(they invented it)
Sure, but AMD has decades of X86 experience, too. They were, after all, the second source supplier of the the Intel 8286, the AMD version being called the Am286. And AMD did, after all, invent the x86_64 ISA that we're all now actually using, while Intel was over fucking around with the utter failure that was Itanium.
Afaik AMD improved licensed Intel 286/386 designs as part of second source foundry agreement for IBM and their first attempt at their own architecture the K5 was really not that good.
They then bought a new chip startup called NexGen which became the K6/K7 later the Athlon.
I think it's a combo of Intel failing with 10nm paired with Apple bankrolling TSMC. Apple + TSMC having a deeper pocket than Intel.
AMD benefited by TSMC basically already investing in smaller nodes for Apple, combined with new manufacturing innovations caused by their innovative chiplet/infinity fabric designs. This enabled them to glue more cores together at higher yields and therefore winning massive benefits in multicore performance.
Intel had to worry about constraints caused by both the architecture combo-ed with manufacturing. I'm also not sure if they were even looking at something like chiplets at the time.
There's a lot of truth to this, but the TSMC part of the story isn't entirely true: first-gen Zen was pretty competitive already on GlobalFoundries' process, at least on multi-threaded workloads. Of course the real trouncing of Intel has only started after they moved to TSMC, but the design knowhow was clearly there before.
Would the significant of Intel's current x86 bread and butter (core 2 duo line and descendants) being essentially a side project of their Israeli Labs, while their main design efforts, i.e. Itanium,etc, floundered a decade ago be an early harbinger of their demise?
Tbe way I see it the problem with Intel is that while they had a great head start, their total marketshare in the fab business was eclipsed by TSMC because Intel didn't for the most part ran their fabs only for Intel. TSMC has thousands of customers pouring money into the company and R&D efforts while Intel was only using them to manufacture Intel parts. They pigeonholed themselves, which is why Intel inevitably fell behind in R&D.
Intel were top dog for several years, and were able to continue making huge profits while only making small, incremental improvements.
End users lamented the lack of real improvements, but Intel continued to sit on their laurels because it was the easy thing to do - they simply didn't need to do anything drastic. Which it pretty sad - with different leadership with technical vision, who knows where computing power would be today? But that's business...
Intel took it too far though, and didn't seem to see AMD catching them up in their rearview mirror. IMO it serves them right, and I'm so happy there is a underdog challenger that is outpacing them in so many areas. The desktop I bought last year has an AMD processor - the first in my household in many years, but hopefully not the last.
They grew fat, probably acquired a lot of middle managers not contributing. And they were so far ahead they were not thinking about the customers and innovation but doing market segmentation and squeezing out more money. They thought they were invulnerable.
They have decades of X86 experience(they invented it) and literally more money than God(they probably spend more on office supplies than AMD on R&D).
How do you have everything going for you and still lose?