10 years ago, I argued that RISC OS needed to focus on small portable devices, but I think it is too late now -- other solutions are now sufficiently advanced that there is no definite advantage to using RISC OS.
RISC OS might have been useful in an OLPC-like initiative, but that is also too late now.
So, I can't really see RISC OS ever becoming mainstream. At best, it will live on for a while as an enthusiast minority system, or various parts of it (such as the font system) might migrate to other platforms. Or it could have a use as an embedded OS (used in GPS navigators, MP3 players, HD recorders, or other specialised systems).