Another point is how Apple hasn't provided other companies with early access to the platform in order to release their Software at launch such as Avid, which is an audio production platform predominantly used by Apple users in Post-Audio for music, tv and movies.
Citation needed: Adobe, Microsoft, Parallels all had native products shown at WWDC so Apple clearly provided them with early access, and native versions of those products started appearing a few weeks after the launch. Lots of other developers took advantage of the developers system and had native versions out very early. Parallels were at a disadvantage because the A12 in the developer's system specifically couldn't do virtualisation, but even they had a preview out after a month.
More likely, Avid can probably afford to take it slowly and wait for the higher-end Apple Silicon Macs, because whereas MS and even Adobe have a significant number of "consumer/prosumer" customers, I suspect that the typical Avid user
isn't so likely to rush out and buy an entry-level Mac.
It's an amazing first step but won't replace a regularly bootable Windows 10.
Except (once Parallels have got the bugs out and done a deal with MS) it
will meet the needs of many users who just need Windows to run a few minor tools that they don't have Mac equivalents for. Virtualisation is more than good enough for anything from running the Windows version of Excel, MS Project, MS Publisher etc. to modest 3D gaming, and has many advantages (like no rebooting) for anybody using both Windows and MacOS side-by-side.
The people left out are the ones who choose Macs for running demanding Windows games, creative and scientific software that are only viable under BootCamp. Frankly, I'd anticipate that most of those people won't be happy with Windows on ARM, even under BootCamp.
Remember, this is all about
numbers not simply whether BootCamp-dependent users
exist or "doesn't work for me". Apple is always happy to throw customers under a bus for the greater good.
Some of those users will only be served with x86 Windows - and that ship has sailed, that need can't be served without throwing away the advantages of Apple Silicon. Some of those users will give virtualisation another try and be satisfied. Some will switch to PC-in-the-cloud services (...and some who currently aren't allowed to do that "because security" are, in the near future, gonna find that they are
required to do that "because security" just as soon as their employer outsources their security/privacy compliance to MS or Amazon). The demise of Internet Explorer has decimated the need for testing web apps on Windows... Since 2006, when BootCamp was a killer app, the range of software for MacOS has improved, the use of platform-independent web technologies has mushroomed and - if all else fails - you can now buy a fairly powerful NUC, Surface Pro or other ultra-portable PC capable of (relative) performance that would once have required finding space for a full-sized desktop or tower. The need for BootCamp - and Windows in general - is a shrinking pool.
Not as simple as you make it out to be. Microsoft's own licensing terms for Windows on ARM restrict it to OEMs who make machines designed to run Windows.
All they need to do is cut Apple (or, more likely, Parallels/VMWare) an OEM license, so you'll be able to buy a version of Parallels/VMWare with a bundled download of WoA. Licensing OEMs is one of MS's core businesses so that's only a stumbling block if MS decide they want to throw a spanner in the works.
Whether there are any
technical hurdles is still unknown: it's no big surprise that a tech preview of Parallels running a tech preview of WoA is flakey and incomplete right now, but there have been some suggestions that some of the bundled apps in the current version of WoA are still in ARM32 code which won't run on the M1, which
would need a fix from MS. However, that's probably on MS's to-do list anyway, since future non-Apple ARM chips will want to ditch 32 bit ASAP as well (or risk turning into the new Pentium 4).
Apple is on the record stating that they are waiting on Microsoft, as Big Sur and the M1 processor can run WoA without issue.
They're also on the record stating that any support for alternative OSs on Apple Silicon will be via virtualisation and not direct booting.
As far as we know, WoA
won't direct boot on M1 unless Apple releases a lot more information on how to write bootloaders and drivers for M1 graphics and other functionality that
did use generic/third party PC hardware but is now strictly proprietary. (Or, more realistically, Apple would have to write Windows drivers themselves). That's a pain for Apple because, at the moment, they don't even have to maintain a stable hardware interface, as every new Mac release can coincide with a point release of MacOS with updated drivers. I don't see MS reverse-engineering the M1 to write unstable drivers that would break with every new Mac (if not every firmware update). There
is someone now proposing to do that for Linux - best wishes to them, but I'm not holding my breath.
Running under Parallels, though, doesn't need M1-specific drivers - the hypervisor either emulates something the OS already supports - or installs a paravirtualised driver written by Parallels - that just passes calls to the hypervisor - and translates it to MacOS framework calls. The fact that the preview is turning in decent performance shows that job is well in hand.
So, WoA BootCamp is most likely not going to happen (although anything is possible) while WoA virtualisation is already way past the proof-of-concept stage and will probably happen unless MS decide to be obstructive.
What is slightly ironic is that the launch of the M1 Macs has done an awful lot to boost the credibility of the idea that there might be life without Intel and that Windows for ARM could be a contender. If it hasn't done already, the M1 MBA is going to be outselling the existing ARM-based PCs, and even licensing a virtualised version of WoA would give MS a huge boost in WoA adoption so, really, why wouldn't they?
OTOH, the pressure is now on the PC/Windows/ARM SoC industry to stop faffing around with the sort of so-so processor that the Surface X uses and produce a proper ARM (or even RISC-V) chip properly optimised for running Windows/DirectX etc. the way the M1 is tuned for MacOS/Metal. That would probably be easier without
direct competition from Apple Silicon running Windows on bare metal - although they ought to be able to beat it running virtualised, so that could be the best compromise.
