Hah.
You don't understand how transitions happen in PC compatible-land vs in Apple land.
In Apple land, Apple launches new thing, provides an emulator that makes new thing compatible with old thing for a while, then people adjust their software to the new thing. Then compatibility with old thing potentially goes away. That's how they did the PowerPC transition, the MacOS X transition, the Intel transition, the Apple silicon transition.
In Windows/PC land, people buy the new chips to run the old software, which it runs better than the old chips while the new abilities are unused. 98% of 386s and a good chunk of 486s went to the e-waste pile having never run a 32-bit protected mode operating system, only DOS/Win3.11. The first ~5-7 years of amd64 chips ran 32-bit XP, 32-bit Vista, and maybe 32-bit Windows 7 and never ran a 64-bit OS before going into the e-waste pile. Etc.
Every time Windows-land has tried an Apple-style transition (Itanium, ARM Surface, etc), it has been a colossal flop. The one transition that was a success, from DOS/Win3.11 to 32-bit NT, it took about eight years and Microsoft had to engineer an entire OS family (Win9x) the purpose of which was to provide enough compatibility with both DOS/Win3.11 world and NT land that people would be able to transition slowly. And until XP came out and forced the low-end-hardware vendors and game developers to support NT, if you were in NT-land, you had to be careful buying software and peripherals because a lot of vendors did not support the NT-family OSes.
And there's a simple reason for this. In order to do a transition, you need four things:
1. The new hardware.
2. An operating system compatible with the new hardware.
3. Drivers, both for things inside the system and for things that people will plug into it, and
4. Application software.
If you look at Apple's transitions, Apple can deliver #1-3 on launch day. If you are on March 14, 1994, picking up your shiny new Power Mac, you have the hardware, you have an operating system that runs on it, you have drivers for whatever video/audio/SCSI/networking/etc componentry is in the thing, Apple has provided enough compatibility with third-party peripherals you care about (e.g. the printers that work on your Quadra will work fine on your Power Mac). The only thing you are missing is #4, and Apple has provided you with an emulator that is 'good enough'. Same thing if you are picking up your Intel MBP on January 10, 2006. And because Apple can start building an installed base on day 1, the Adobes and Microsofts and other third-party developers look at this and say "okay, this new platform is happening, time to start recompiling our code guys".
If you look at PC land, ever since IBM stopped setting the standard in the 286 days, different people are responsible for each. New processor architecture is released on a given day; that's nice, there aren't going to be any operating systems. Microsoft may recompile NT for your architecture on whatever timeline they feel like. Linux folks may compile their world for your architecture, again on whatever timeline they feel like (and just because the kernel boots on your thing doesn't mean that the established distributors like Ubuntu/Debian/Fedora/etc will have a distribution for your architecture for a few years.) NetBSD folks will probably port their OS to your architecture; other BSDs may or may not, and even if they do, you're probably not going to be a "Tier 1" (in FBSD speak) architecture for a few years at least.
Then you need drivers. No one is going to be coding drivers for printers and scanners and network controllers and GPUs and whatever else for a platform that doesn't exist. So, on the day that Microsoft ships NT for your architecture, congratulations, you can't plug any hardware into it. Then, of course, you have the software problem, but your OS vendor can at least provide an emulator for that. Maybe. And all the third party software developers are watching this clusterf*** of a transition that's going nowhere and they think they probably should wait to see how things flesh out before they assign developers to porting their software.
If you want to see this in action, look at how many developers support i) 64-bit Windows, ii) ARM Windows, and iii) Apple Silicon. You can barely get 64-bit Adobe Acrobat for Windows/arm64; the same thing has been available in 64-bit Intel for macOS for years and in 64-bit Apple Silicon for a while. Is there ANY third party software for ARM Windows, other than maybe some open sourcey things like VLC that get compiled for everything? And actually, VLC doesn't have a release version for ARM Windows, just nightly builds.
Meanwhile, with a kludge like AMD64, AMD releases the processors, they keep running 32-bit XP with existing drivers, software, etc. Which they do better than the other x86-compatible processors on the market. A while later, MS releases amd64 "XP", which no one uses because there is almost no driver support for it. But at least it exists and MS indicates that amd64 is going to be a supported architecture going forward. Vista/7 deliver equivalent feature sets and experiences on both x86 and amd64; by the launch of Windows 7, there are enough drivers for amd64 that consumer machines start to be preloaded with amd64. Meanwhile, businesses are still cautious - my dad, for example, got a new work laptop in 2011 with a sandy bridge Intel that was running 32-bit Windows 7. That's eight years after the Athlon 64 shipped. By... 2013-2015, amd64 is established enough that businesses are now abandoning 32-bit. And it is only in 2021, with the launch of Windows 11, that Microsoft stops shipping 32-bit Windows. So... eighteen years.
If anybody thinks that Windows-land and the established base of Windows software can move away from x86/amd64/etc, they are dreaming. More likely to move to Chrome OS on arm than Windows on arm. Windows on ARM will never get off the ground because... you have no installed base to motivate people to write drivers, no installed base to motivate people to port their application software, and you have nothing to motivate people to buy the devices so you can't build an installed base. Meanwhile HP/Dell/Lenovo keep shipping millions of amd64 devices every month. Microsoft couldn't even get touchscreens added to the Windows platform when they went big on touch with Windows 8 - Lenovo/HP/Dell basically said "no, we don't think people want to pay for touchscreens", 80% or whatever of Windows machines they shipped had no touchscreens, and... here we are, a decade after the launch of Windows 8, and touch on Windows remains a joke.
And Microsoft may be "dead serious" about Windows on arm. Just like they were "dead serious" about a smartphone platform when they realized that smartphones were going to muscle in on the PC. You can be as "dead serious" about something as you want - if you can't ship a functional product that enough people are willing to pay for that a third-party ecosystem builds around your platform, you are doomed.
And Windows on ARM has even less of an ecosystem than Windows Phone...especially since Windows Phone had at least somewhat of that Microsoft mystique from the 90s (people like my dad, when he got his first iPhone, expected that Microsoft would come out with a phone and everyone would switch, because, well... that had been how things had gone the previous couple of decades - Apple invents something, Microsoft copies/"improves"/mainstreams/out-markets/etc it, everybody switches to the Microsoft version. The idea that Microsoft would enter a market pioneered by Apple and... completely flop... and end up exiting in shame was... not really conceivable before Windows 8/Phone.)
