Actually ARM based devices that are going to run full Windows 10 won't need any special modifications of the centennial apps because of the emulation.
Perhaps we shall see what is their plan for phase two in the upcoming events.
It is not going to help mobile directly, but it's definitely going to make Windows Store more populated with quality apps, which they desperately need. That should show developers that the Windows Store can be profitable, so perhaps then they will start "converting" their Win32/Centennial apps to full UWP apps.
IoT core, HoloLens, Windows 10 mobile and console all require UWP apps not win32, and three of those are ARM based. At least some of those will not be running full windows 10 in any kind of quick timeframe due to hardware requirements, OS function, and to a lesser degree the construction of the shell.
Take something like the Kodi centennial bridge app. It runs great on a big screen, and works fine on a larger tablet, but it would start to get a little cramped on a smaller tablet, and wouldn't work on a phone sized device. It's "somewhat" touch optimised, but not fully, doesn't scale well, doesn't run on xbox, or phones, or IoT core or HoloLens or mixed reality.
I think like the software bridges, windows on arm, is like an "os bridge", to help the legacy uses move over to the modern way of doing things. Once most things are UWP, its served its purpose.
But yeah, as you say, if MSFT can get developers into the store, they will probably want to create full UWP apps over time, so they can roll their software to console, mixed reality, IoT, phones, HoloLens, and so they scale well to different screen sizes and inputs for tablets as well.
The genius of the arm move, is that LTE/telephony/GPS enabled tablets and hybrids will also benefit from all those mobility apps - things like snapchat, lyft, banking apps and so on that are missing from the UWP platform.
And THOSE developers won't be coming in via the centennial bridge - they would come in through via islandwood (the porting gateway from iOS) to xamarin 2 - a revised codebase that enables an apparently large amount of code sharing across android, ios and uwp, whilst preserving UI features. For a fair bit of initial work in the port, those coders save work later on, in maintaining three OSes with greater ease.
Resulting in, once windows on arm comes to tablets, fingers crossed, a bunch of mobility focused full UWP apps of the more "smartphone" variety.
Between that (windows on arm/islandwood/xamarin 2) and the centennial bridge/window s, all things going well, the windows store will look considerably healthier by the end of next year in terms of scaling full UWP. Healthy especially because it will be gaining from desktop apps, not merely mobility apps.
I know people complain about the store, how long things take, but the scale of what MS is trying to achieve is bigger than anything its really ever done. The shift to 32 bit and windows 95, or NT/XP is minor by comparison. And they have to move developers and users into the new way of thinking in terms of modernizing the app platform and creating a unified platform.
It's a huge task. Even what MSFT themselves are doing at the OS end is mammoth (trying to get windows 10 to handle, virtual reality and mobile, and wearables and all that adaptively).