Brian recently posted a really interesting discussion about hosting Windows on Azure and the barriers Microsoft puts up. I see two issues at play here. The Microsoft client business model, and Microsoft's definition of VDI.
VDI is not just about a desktop or a server operating system
Microsoft defines VDI as a Windows Server role and calls it RDS, which is really shared hosted virtual desktop (SHVD) that runs only on a server operating system. With the previous Windows Client regime, VDI (as a hosted desktop operating system-Windows XP/7/8) never stood any chance of succeeding. Sinofsky never gave a crap about the enterprise and was laser-focused on his "me too" Apple compete. Let's also not forget the failure of Microsoft to bring the phone and desktop closer together earlier (like Apple), versus the cluster f*** that is Metro and Surface tablets today. Even if there was a voice for, "Hey guys make VDI (desktop OS ones) work," it was a tiny percentage of the market so nobody cared. That's the problem with Microsoft—they only care about big numbers and don't build the products (especially for the enterprise) that we actually may want and take a chance that they will grow to be big. This leads to them being out of touch with where the portions of market are going. The fast follower mindset in a world where things are changing rapidly is a failing business model.
Obsession with device CALs needs to die and evolve faster
Microsoft is still about protecting their client device CAL business model and brain dead sales force who are robots that will defend this until they are forced to change. If they were smart, they would understand that a percentage of enterprise desktops are going to a VDI and SHVD model. This is going to happen because it solves real problems. So with a new regime, my hope is they will understand that they are losing appeal in the enterprise, and this is one small thing they can do to keep people on Windows. The recent price increases demonstrated that they do understand user-based licensing, even though people are bitching about it. Perhaps this is all a function of time in a huge organization. All Microsoft is doing is hanging onto a legacy device-based business model until they figure out how to change it without blowing up their cash cow. I hope that is the case versus they don't care, since then Ballmer justs needs to decide "let's fix this problem before we lose yet another market," win goodwill and just make it happen. IT'S THAT F'ING SIMPLE!
There's an opportunity to make Azure more interesting sooner
However, due to their definition challenge, even if Microsoft did something with VDI on Azure, I have no doubt that it will be on RDS which they will claim is VDI. They will say we have a SPLA license model for rent style use cases, they will say it's cheaper (not the use case to do VDI) and they will say it's good enough for the mass market. This would be the typical response of a monopoly that is too blind to see the ship is about to hit an iceberg and then wonder why customers still remain frustrated.
I'd like to see RDS on Azure for many use cases. If VDI with a desktop OS was also available as an Azure workload, I may actually stop thinking Azure is not a serious player for many years to come versus the rest of the ecosystem. It may even start to make sense to consume apps written on Azure for desktops hosted on Azure and do all sorts of interesting stuff with Windows tablets. The question remains, does Microsoft have the vision and understanding? I am sure they have people who get it. Can the client and server groups work together to solve a problem that the world wants them to? With Sinofsky gone, there's an opportunity to change and demonstrate that collaboration in a monopoly is possible, since lack of collaboration has been touted so much as his downfall.