IIRC, the selling point vs. Steam on Windows is that somehow Windows OS is holding back performance - so with a custom Linux OS, theoretically games would perform better. The idea is that by putting the weight of Steam/Valve behind it, there may be enough critical mass to force a) hardware manufacturers to solidly support the Linux drivers and game developers to actually release the games in a Linux compatible format with c) good controller support for the PC (no keyboard required.)
It will depend on gaining a sizeable market share fairly quickly in order for it to gain critical mass; based on the skeptical reception so far, I don't really see that happening.
Personally, I don't think it's the right problem to solve - PCs generally have far superior performance to consoles at any point other than immediately after the consoles' introduction - with or without Windows. Also, game developers who target consoles are still going to have to develop for the other major contenders, even if they also target SteamOS - so it won't "unleash" some super-performance game that the Playstation/Xbox consoles only get to play in some hamstrung format. Lowest common denominator still wins.
On the plus side, all of these Steam boxes will be fully capable of running Windows or other Linux distros should the SteamBox thing not work out (vs. say a PS4 - which if Sony went away tomorrow would become a big paperweight.)