The Mac market is not the Windows market.
If MacOS is better than Windows, then one of the key reasons is that Apple can have a clean sweep every decade or so without losing their ultra-conservative corporate customers. Even Apple's "pro" users mainly work in media creation and, by necessity, are vastly more open to change than the banks and insurance companies who keep Microsoft stuck with supporting 30 year-old software.
Windows users expect to be able to run binaries compiled in the 1990s (and they have to support a huge corporate market who rely on this) using an API that was x86-centric since it was cloned from CP/M (...Windows NT is newer and cross-platform, but the non-x86 versions all failed and/or were dropped, and even Windows 10 still supports old Win9x-era, if not before, binaries).
In that time, Apple have (a) completely switched processor architecture 3 times (68k to PPC, PPC to x86-32, x86-32 to x86-64) and (b) dumped Classic MacOS for a completely new NextStep-based OS that isn't even source-code compatible with "Classic" (and the "classic" emulation mode is long gone)... which, as a Unix implementation, is fundamentally platform-agnostic and founded on source-level compatibility anyway
Also, Apple control the hardware and the software, so they can force a transition by simply announcing an end-of-life for x86 Macs - anybody wanting to stay in the Mac software business will have to support ARM - whereas Microsoft can't (well, they could drop x86 Windows... and cease to exist) so currently you have 1 or 2 ARM Windows machines vs hundreds of Intel-based competitors with no particular incentive for developers to support it. Heck, Apple could afford to lose the Mac entirely.
Meanwhile, have spent years shepherding developers towards official MacOS frameworks for graphics, acceleration etc., depreciating OpenGL, CUDA etc. They just purged the Mac world of "abandonware" and legacy binaries by dropping 32 bit support in Catalina. For the vast majority of applications, written mainly in high-level languages, recompiling for ARM64 is going to be a trivial job compared to converting them from x86-32 to x86-64... and for the exceptions, well, if the developers don't think their worth supporting they're likely to die at the next major MacOS upgrade anyway (if they haven't already been killed by Catalina).
It's pretty ironic that the most serious loss from switching the Mac to ARM would be the ability to run Windows
...although, personally, that's a feature I'm finding less and less important - increasingly, there are Mac and cloud-based alternatives to Windows-only software, the demise of Internet Explorer and MS's switch to Chromium vastly reduces the need for testing websites/webapps on Windows while the lack of a touchscreen on the Mac probably means that I'm going to end up needing to buy a Windows laptop/convertible to properly test websites/apps.
Meanwhile, if I really need an x86 Linux VM (given that ARM64 Linux is already well-developed), I can spin one up in the cloud for a penny an hour (its not like you can do modern web development without an Internet connection).
That seems most likely - 6 months of a developer program should ensure a huge swathe of natively-compiled ARM apps. The big pro media apps will likely take a bit longer.
However, notwithstanding the comments above (and the revelation that an iPad Pro+ Magic Keyboard weighs more than an Air) I'm starting to wonder if the mythical ARM Mac will actually turn out to be an "iPad Laptop"... which would explain all the interest in "Pro" apps for iPad.