MS is going to make damn sure WinRT succeeds. 4 million people upgraded to Win8 over the launch weekend, no telling how many have since then. We don't know how many Surfaces have been sold, but we know certain SKUs are sold out.
Those 4 million people who upgraded to Windows 8 can also run Win32 apps just fine.
I don't know how many bought a surface, but it isn't hundreds of millions, which is the sort of number you need to for WinRT to start looking more attractive than iOS or Android, much less Windows Win32 API itself.
Windows Store is put front and center right in their face with apps at their fingertips. If you have a competitor with their app in the store, the are -bar none- getting better access to your potential customers than you are (even if you get a certified Desktop App link in store, they still have to jump out of store to get it and therefore a higher friction to sale). There is no way in **** I'd let my competitor have better access to 4+ million customers just because I don't want to write in WinRT.
For now I can live with that. There are also hypothetical competitors beating me for a few Mac OSX customers, as well as for some BeOS customers, not to mention those millions of iOS and Android customers I might miss out on if I waste time chasing those handful of Surface RT customers. I can't target every platform. I have to pick and choose the ones that will get me the best return, and position me for the future. At the moment a WinRT app seems very unlikely to have any significant marginal returns anytime soon, and it's future is much less certain than you seem to believe. If WinRT picks up, then I'll reconsider.
I could be wrong about all of this. Feel free to pick up Visual Studio and start working on some WinRT apps.