This might sound ominous, except for the fact that Valve seems to be taking a long-term approach. Valve has emphasized repeatedly that it is interested in an open alternative to Windows and OS X, and this interest might even survive the failure of Valve's current plans.
However, the success of Linux gaming does not depend on Valve alone. The slow development of Linux gaming suggests that the development companies have failed to understand their market. Perhaps they have underestimated the size of the potential market, or its possible appeal to Windows users who are already using what -- to them -- is a perfectly adequate operating system for gaming.
In particular, they may lack insight in the psychology of Linux users. The first figures for Linux users on Steam suggest that the potential audience was curious, but was unimpressed by what they found -- almost half did not become regular users. Perhaps they were disinclined to pay for games, or disliked digital restrictions management, although these complaints have largely gone unvoiced. Or perhaps Linux users are simply disinclined to participate in manufacturer's polls. However, whatever combination of factors is involved, it seems that you cannot sell Linux games the way you sell Windows games.
Yet another possibility is that commercial Linux games are happening too late. Although Steam can offer the popular games, a small but growing number of free-licensed games like 0 AD and Unknown Horizons are starting to emerge, many of them featuring contemporary graphics and game play.
True, these games seem to have small audiences, but they are more in keeping with the free software ethos than any commercial game. Give them a few years – commercial games may have trouble competing with them.
To say that commercial Linux games are a failed concept would be premature. However, we can now say with certainty that their success depends on that of the Steam Machine, and, perhaps, the manufacturers re-examining their approach -- all of which should make for an exciting next couple of years.