Totally unofficial, subject to complete change, take with a gain of salt.
The following seem to have bright futures regarding hardware decoding, from my fly-on-the-wall perspective:
AMLogic devices - this one is obvious because we get support for some AMLogic devices directly from the core Pivos work, and that could spill over to other AMLogic devices (technically already has).
Tegra 3 devices- Since we said we would support Ouya, that could spill over to other Tegra 3 devices as well. Also, almost half (or more) of XBMC's devs seem to have Nexus 7's... :)
Exynos devices - Samsung's chip, and what you'll find on the very awesome ODROID series of dev boards (that are cheap enough that they'll probably make for great ARM HTPCs). Not only are their CPUs powerful enough to handle software decoding of some files in 720, but there's some early work being done with hardware decoding and there's already some proof-of-concept XBMC builds that hardware decode h.264 in 1080. This could spill over to other Samsung powered devices.
There's work that's been done (not sure if it's still being worked on, or if abandoned, as far as Team XBMC goes) on Allwinner A10 devices, but the libraries that Allwinner provided have tons of bugs. Last I heard was that things were progressing at a snail's pace as far as getting help from Allwinner, and our devs were basically strung around by them in the past. I get the strong feeling that Allwinner A10 is a dead-end for XBMC as of right now.
There's also some "catch all/some" work being done that will work on multiple chipsets with something called libstagefright. When you look at the Pull Request (
https://github.com/xbmc/xbmc/pull/1832 ) you get a tiny window on what just some of the issues that are faced when any of these solutions are attempted. To me, this seems like a very exciting development, but I also don't understand most of what the devs are talking about in that discussion :)
I'm not sure, but I also believe that something called OpenMax is also another attempt for "catch all/some" work.
I suspect that any ARM/Android devices that use Broadcom hardware video decoding might be able to take advantage of what we know from the Raspberry Pi, but I could be completely wrong about that (total speculation there).
These are just the examples that I can think of off the top of my head. The devs also work on things in their own code repos and only bring things up when they're at a point where they either need to collaborate with other team members, or enough work has been done to merge, or something like that. For example, I didn't even know about the Exynos work until after the ODROID people posted a youtube video showing XBMC playing 1080 videos on their hardware. So there's various things at various stages.
As always, we don't want to get anyone's hopes up, and more importantly, we don't want to be responsible for anyone buying something that turns out to later not work (outside of the few devices that we've said will work, which I think are just the Pivos XIOS and Ouya). Just know that there is very active work being done on this, and the end result will likely make for a great selection (hopefully "most") of Android and/or ARM HTPCs.
EDIT: And for those who are wondering about why other media players are able to work with hardware decoding already, but XBMC doesn't have it yet: XBMC is an open source project and any hardware video decoding code that's directly inside of XBMC also needs to be open source. We also can't use stuff that requires us to withhold code do to Non-disclosure agreements (and even if we could, GPL wise, I don't think most of our devs would like that). Most of what you see that works with hardware decoding involves stuff with NDAs or some kind of closed source code.