1. Microsoft does have a bad reputation and has many fails where they have gone from the biggest to nearly nothing. I did note that this is probably the closest that anyone has come. I just don't see it. Windows share won't decline, it can't decline. Mainly because you have to include DX11 support in the software for a long time. The chances of seeing someone do DX11 and Vulcan instead of DX11, 12, and Vulcan is going to be minimal.
2. There has been. Hell for 5 years everyone was using the either Quake 2 or Quake 3 engines in just about every game. Those were designed around OpenGL because Carmack didn't like and didn't trust Microsoft. Hell 3d gaming started on OpenGL/Glide. id was making so much off of their engines that they basically took like a 6 year break from developing any games. 2/3 or 3/4 of the AAA games that came out around this time had deep support for OpenGL, but almost none of them actually even enabled OpenGL support in the software. Now back in the day it was much smaller teams with much less money to spend. But still it was there, but ATI cards sucked in OpenGL, and it was better and cheaper to work with the one API that worked well with both.
3. Nintendo is another beast altogether. On one hand the Switch might have sales numbers to push companies to develop for it. On the other hand it is still so far behind Sony and MS (even their 4 year old models) that it needs it's own development branch. Capacity limitations means it requires compressed assets. Compressed assets means more CPU overhead load assets. Much worse CPU and GPU with a much smaller feature set. Non-compatible CPU. Nothing that is done on the switch is really transferable or vice versa.
4. Mobile market is the same as the Switch and this is the same scenario that they have been facing since Apple first launched the iPhone. People developing for OpenGL for platform specific offerings still don't go back to PC development and start writing for OpenGL in desktop games. In fact even if the mobile market game is Rockstar, EA, or Activision, they are written by teams that work primarily on mobile market games and don't have anything to do with desktop development. It doesn't matter to EA they have to OpenGL on a play store app and DX12 on the desktop. What EA cares about is the most amount of sales with the least amount of development time. Since mobile apps are their own monsters, then desktop development will have a completely separate eval.
So the question then becomes. And this ties back on to your feelings of Scorpio (and god that sounds smug talking about 24 million consoles like its bad thing.) It's not just about exclusives though its the most important part. It's about interface, its about connectivity, its about controller feel. I have a One and PS4 and I like my Xbox a lot more for gaming so the only PS4 games I get are exclusives. Which only about 3-4 really good exclusives to me are even available. Sadly the PS4 spends a large portion of it life powered off. But I am probably an edge case. Scorpio though will be a different beast. The Xbox ecosystem already established. The resources Scorpio will offer will allow for easier porting over from PC or vice versa. The Xbox uses DX, and the GPU will be a DX12 GPU. So a developer can work on optimizing Vulkan on the desktop and then DX on the Xbox? It's not going to happen every game that will be supported on both will optimize both for DX. Why because even with the Xbox's "lame share" game sales on that console are much more important than PC sales. We know console development drives PC development and it's useless to say otherwise. Some developers also try not completely screw over PC users, and we can thank them for that. But if Vulkan is only a neutral to slight positive move from DX12 there is little reason to optimize the game for it even if the engine supports it. That is where this RotTR video and subsequent DX12 on Nvidia hardware eval is going to bite Vulkan development in the long run. We are still a while away from seeing Vulcan options in game as a normal thing. Companies will hedge their bets for a little while by including DX and Vulcan support in games. Having AMD with working DX12 options if combined with a competitive card will pressure Nvidia to optimize their hardware/software for DX12. The closer that DX12 comes to having a truly optimized performance on all Nvidia and AMD hardware the higher the hurdle Vulkan has to overcome. If Vulkan doesn't have an actual performance benefit than the little amount of hooks and calls in the PS4/PS4Pro/Switch when writing for metal will mean nothing, when they have do the same for the metal code with DX hooks on the XboxOne, and the much more API driven implementations on Scorpio will require DX and there is little to no reason to duplicate that work if they can just copy paste it into the PC version since there won't be a performance benefit.
Scorpio doesn't even have to sell that well. But it brings up another point. A lot of PS4 sales growth and continued growth was due to it having a visible but mediocre performance advantage over the Xbone. Scorpio will have a massive lead if rumors are correct. I have to think that it will impact Xbox adoption to some degree.
Jeez, this replying is starting to take time...
1. Windows share is declining and will decline. Windows server is generally lost the battle in its territory. On workstations and desktops windows will be for a some while, but as more and more services are going to cloud the relevance of used os is diminishing. We already see often Apple in this sector. Linux is slowly but steadily growing and Windows is at the same time struggling with its W10 as there are little to none reasons for upgrading from W7. If you watch the near future in wider than games perspective, it's hard to find any logical paths which would lead growing windows share or even staying the same. World is changing, os relevance is decreasing and unluckily for Microsoft it's not their way. That's why they are desperately investing to all other areas now: Azure, Office 365, competition with Steam, merging Xbox and Windows as they still have the advantage with DX. However, if they don't open their operating model a lot, they're eventually going to lose this game. This has been the overall trend in business computing for a some while and the same is happening in end-user segment.
2. Time is now different. OpenGL is matured and globally used in variety of platforms. Vulkan is mentioned already in nearly every place where there are discussion about dx12 and future games.
3. I wouldn't be that black-white with the assumptions of Nintendo ecosystem. Remember that Nintendo has already expanded to mobile platform which also has the "compressed assets". Point is that Nintendo has made truckloads of money with their two mobile games hence underlining the fact that mobile gaming market is very noteworthy player. Switch also nicely pinpoints the idea that you can have mobile and console in the same packet. Sony had something that you could play PS4 games on your phone, the newest S8 can also be used computerwise and most likely more to come. In here you need to have an open API, not some MS driven proprietary solution.
4. This is also changing. Of course there are differences with the user interface (no mouse, keyboard etc), but not all graphics, logic, physics, audio etc. need to be different anymore. It's pretty amazing what the current high-end mobile devices can bring to the table. As you said "most amount of sales with the least amount development time" means that there are urge for common interfaces, languages, tools, APIs in this sector. In this view, which is the actual need on the market, MS dx12 is just a nuisance, Xbox is only place where you're forced to use it. DX will always be MS ecosystem product and MS will never own the whole gaming sector. There is real need for Vulkan type of open, efficient, multiplatform solution.
The major point here is that gaming used to be windows related stuff, but not anymore. There is now way more platforms to support, you cannot ignore any of the big ones but at the same time you have limited amount of money to do the ports. Nintendo Switch, PS4 mobile gaming, Nvidia online gaming, VR solutions, mobile, SteamOS are all here mixing the pot. Vulkan is only one step to have some glue between all these. Everybody knows that MS has their own agenda with dx12, that is simply to keep things in their hands which is not what markets really want.
About owning consoles, I have had all Xboxes so far, but it's still only a console despite all the new features MS have tried to bundle in it. Nowadays it's in very small use, I even use my Wii more with children than Xbox. If I want to play something by myself, I choose PC because I know that I have and will always have better graphics and user interface (mouse+keyboard) there. Also buying games from Steam sales saves a lot of money. All my "console race" friends tend to have PS4 to which they seem to be pretty satisfied. Vast majority of households chose one of the consoles, not two or three. PS4 pro was announced little while ago, VR after that + PS4 has been leading the console war right from the beginning means that people have lots of PS4 games on their libraries. It's really just not the performance which leads to the purchase decision, remember Wii, Microsoft's strict used game politics which was the real reason why Xone lost the battle at the very beginning? Also have to mention that even there are all kind of consoles/playing platforms available, my children and all their friends are using majority of their gaming time with mobile games (clash of clans, pokemon etc). They are the future anyways and maybe that's what Nintendo is thinking with their crossover Switch. Lastly if Scorpio will be way faster than PS4 we can be pretty sure that Sony has a new version waiting around the corner.