Well, that depends on how important Parallels Desktop is to their business (AFAIK their other line is the Plesk Linux containerisation software for web virtual hosting - which is probably also getting decimated by the more modern cloud providers). There's no guarantee that there is ever going to be a dominant, open-ish, ARM platform that would make a generic Windows retail license worthwhile).
\
The WoA system builder market doesn't have to be dominant (over the x86_64 one) ; it just has to be healthy. Windows has about 90% of the market. If WoA could just carve out a 10-15% of that then it would be larger than the Mac market, but x86-64 would still dominate ( 75-80% of the market). Right now it isn't a robust ecosystem. Qualcomm has mainly tried to pawn smartphone SoC onto the market. As long as it is only Qualcomm trying to sell more radios really isn't a viable ecosystem.
ARM doing the X1 and X2 ( and probably X3 and X4 ) cores isn't an absolute guarantee that it going to come , but it is a necessary precondition. ARM SoC with 4-6 X2's and 4 510's would be a far more viable foundation for a Windows ARM platform and that is just standard library parts (no need for a massive design team to get into the "game"). Those cores are coming 2022. It is only a matter of a vendor to allocate enough die space to put them in a Windows targeted SoC. The only thing needed is some limited tweaks and a fab order. Samsung (+AMD RDNAx GPU) + MediaTek (+ Nvidia GPU). It is going to get past the point where Qualcomm is the processor provider choke point. ( And if the Nuvia acquisition pans out for Qualcomm and they move to a more balanced diversification that will help also, but not necessary. )
Intel cooperating with Windows and making Windows 11 far better at "Big Little" will also something that is going to change the playing field for an rollout. Windows 11 dumping 32-bit as a major target mode and BIOS ( UEFI) and going solidly 64-bit only will have a side effect of further enabling the WoA solution. Microsoft spending tons of effort kowtowing to old , legacy 32bit x86 code is one of the boat anchors that has been holding WoA back. Windows 11 is going to help to largely help to cut those 1990's and 1980's stuff loose.
Will there every be a huge Windows 10 on ARM market? Probably not. The issue not though is that Windows is at an inflection and that likely to going to open a door for ARM and also for Intel/AMD. In the latter case to eject a wide swath of the 32-bit now dubious design moves that are hobbling them in the general PC market place ( especially on laptops. by 2025 not going to be surprising to see those two have a significant subset of their product that has "moved on" to a cleaned up x86_64 only. )
The final contributing factor will be Chromebooks. Similar rumblings of X2-510 SoCs coming that the larger players in Chromebooks HP/Dell/Lenovo/Acer system vendors that they sell models with small tweaks into both Windows and Chromebook markets.
Windows-on-M1 is most likely going to be dependent on virtualization - which leads to the counter-question: why would Microsoft take on Tier-1 support for Windows running on Parallels' & Apple's proprietary hypervisor?
They won't . Folks who buy an OEM license are on their own for Tier 1 support. ( the OEM, which is the person who bought it, is response for supporting themselves. ) . The folks doing it for free ... they won't in that case either in a substantive way .
If there are enough largely indistinguishable vitrual machine, lowest comment denominator targets out there they will , because there is scale. When Microsoft establishes a baselien on ARM with Hyper-V then they won't if Apple's hypervisor deviates too far. Microsoft is a hypervison provider themselves. They aren't going to let Apple dictate to them what a hypervision should or shouldn't minimally do for Windows 11. ( Widows 10 is on a 4 year delcine track now; 2025 de-support. )
...except that's a market segment where Linux (which is pretty well developed on ARM and less dependent on binary compatibility than Windows) already has a significant presence, and a large proportion of users tied to Windows will also be tied to x86.
Unix/Linux was a dominate player in data center market before the super hype "cloud migration" started. It dominates now because it dominated before. That what has happened has larger been moving servers from "private data center , private cloud" to "hosted / secured-public cloud". Those are data center to data center moves.
The move to push folks from dekstop OS back into the remote GUI "terminal" services mode with "even thinner" clients is different. The dominate player on desktop is Windows . The Linux desktop marketshare is a minor blimp.
Chromebooks and Macs are bigger players than that. This is actually a different sub-market segment for cloud services vendors. Not just cloud. Lot of business run "private" terminal services set ups which basically would run better , faster with ARM in more than a few cases if WoA+Office+necessary apps is sufficient.
There's a reason that people talk about the "Wintel monopoly" - DOS/Windows and the x86 have been joined at the hip since IBM "made" them in 1981. Windows isn't going away any time soon - there's too much legacy stuff - but I think the rise of mobile and web technology in the last 10-15 years (which orbits around ARM and Linux rather than Wintel) marks the end of "peak Wintel". Hell froze over the day that MS announced SQL Server for Linux...
x86-64 was pragmatically well deployed by 2006-7. Roll forward about 20 years ( 2026-27) there is a newer pile of now legacy stuff to support. The 40 and 50 year old stuff aren't really going to be a critical part of the business.
That could be largely chucked from most of the respective product lines.
As for SQL Server. Chuckle. SQL Server started off as a fork off of SyBase which was multiple platform in the first place. It was Ballmer , Gates, and Microsoft super ego that made it single platform. Someone managed to take their head out of the their butt and look around isn't "hell freezing over". IBM had mulitiple platfrom versions of DB2 for over a decades before Microsoft did. There was no way Microfsoft was ever , ever going beat Oracle and DB2 with the bonehead path they were on.
