There's something else that happened on our way to Apple Silicon: iPhone.
Intel ruled the CPU world with an iron fist for a few decades, and were able to maintain it because they had 3 key advantages: backward compatibility, process leadership, and boatloads of money to fund an elite design team. This combination has kept them ahead and prevented anyone from seriously considering alternatives. Backwards compatibility has kept the bulk of the Wintel market in Intel's pocket, this funded an R&D operation that could make the best of an archaic and inefficient architecture, and the fact that Intel always had a superior process made up the rest.
PowerPC may have been a better CPU architecture and, if given the budget Intel had, might have continued to outperform x86-- but it didn't run Windows (ignoring the Windows NT experiment of the 90's), so it was never going to get the same revenue.
AMD was first to 64bit when Intel got their ties stuck in the thresher of Itanium, but Intel took the initial setback and just started using their process advantage to ramp up clock speeds until they clawed their way back.
Intel never really addressed the embedded/mobile market. I think they saw it as low margin, low performance. They'd had licenses for some of the best performing Arm variants in the past, but never capitalized on them and sold the IP. They still have Arm licenses that they use for some of the MobileEye and FPGA stuff, but they never worked to establish a beachhead in embedded and that left them with a soft underbelly that Apple exposed.
When iPhone happened, Apple was motivated to build a high performance, low power processor so they went back to what they started with Newton: Arm. But Apple sold a lot of iPhones. Crazy numbers of iPhones. I'd have to go back and check the numbers, but I'm pretty sure Apple has been making more A-series processors than Intel has been making x86 processors. And they're doing it at a high margin.
Lots of processors at a high margin translates to an R&D budget that can start to outstrip Intel.
It has also helped Apple build critical partnerships with suppliers, in particular with TSMC. Building supplier partnerships is one of Apple's great strengths. While Intel keeps stepping on their tail, TSMC has been delivering on generation after generation, so Intel has lost their process advantage. Even Intel is farming work out to TSMC.
And Apple showed us something else: Arm, with some clever tweaks, can run translated x86 code about as fast as x86 runs it native. The backwards compatibility challenge isn't as scary as the industry thought. This isn't a huge surprise, really. The whole reason Apple pulled Arm out of Acorn into a JV was because at the time Acorns RISC processor was able to emulate MacOS faster than it ran natively, and without running hot.
The sheer scale of iPhone development is what made it possible for someone outside the x86 ecosystem to finally compete head to head with Intel on processors. With those three advantages, there's no reason for Apple to continue using x86. Add to it that the ability to create custom logic for system functions and to finally be able to move away from Intel's "everything on the CPU" mentality is a perfect fit for Apple's system level approach.
I'm sure Apple is quite happy to keep on the Apple Silicon roadmap, but I also expect that as soon as something makes it the less preferable option they'll pivot again. If TSMC fails to execute for whatever reason, I'd expect Apple to move elsewhere and to do it quickly. Something may come down the tech highway that Apple isn't positioned to do in house-- quantum computing, for illustration sake, might send them back to working with IBM, or whoever.
This is what I've come to expect. At every moment, Apple will make the best product they can and they'll make it as easy as they can for people to go forward, but they're not all that concerned about looking backward. Carrying forward the burdens of the past just slows down the future.
