I'm an MBA (hold off on the throwing of the rotten vegetables! I'm a IT person too!) So I'd like to put my 2 cents worth on the whole thing from a business perspective.
Everyone's talking about it being a chicken and egg situation where devs aren't making games for Linux because there's no market, and there's no market because there aren't any games. This isn't really the situation. The execs at big companies often deal with situations where they have to take a leap of faith. Every time there's a new console, for example, the execs at companies like EA decide whether or not to make games for it well before the console is released, so they're making games for a market with 0 users! They make the decisions based on a few key factors, including looking at the risks, the chances of success, and the possible rewards given the market. Here are just some aspects that are probably discouraging to an exec at a big gaming company:
1. History. Linux is old. Really old. And it hasn't taken off in the consumer market yet. So it's a pretty big leap for an EA exec to think it's going to get popular now. There hasn't really been any change in the market that would point to a massive upswing in Linux gaming.
2. High potential risks. Xbox isn't that big a risk to support, since it uses similar tech to Windows. Linux? It's a bit different. Sure, it uses OpenGL, like a mac, but it's a whole different platform. This wouldn't be a deal killer by itself, but it's another nail in the coffin since it increases the risks.
3. Lack of proof of a market. As people have pointed out, the Humble Bundles sold well, but they had people giving to them because a. They wanted to support small indie developers and b. they wanted to support the charities that the Humble Bundles give to. When companies look to predict what's going to happen they look for comparability, that is, they try to find similar situations where there was a success, and there is very little evidence for this. Should they take a chance anyway, and do something new? That leads us to the last and perhaps biggest point:
4. Low first mover advantage. One of the things a business looks for is first mover advantage, that is, what kind of benefits do they get by taking the risk of being the first to do something. What they're looking for is some reason to think that going first will let them get and HOLD ON TO a chunk of the market. This isn't the case with Linux. Let's say that Carmack decides to make his latest game (Quake 7, this time it's even Quakier!) in Linux. Let's be generous and say that Q7 is released, the Linux gaming market explodes, and everyone buys Q7 for Linux. Carmack took a big risk. What did he get in return? Well, he got big profits, obviously. But he didn't do as well from this deal as you'd think: Let's say that Blizzard, after seeing Q7's success, produces a first-person Linux game called Starcraft 3D: Raynor on a Plane. Assuming it's of a similar quality to Q7, their profits are about the same. Maybe even better, since the market has now grown even more. But they didn't have to take the risks that Carmack did: they lost nothing by waiting until Linux was already a success. And unlike with a console Linux doesn't have a short life cycle, so they had all the time in the world to wait. It's true that Q7 had the advantage of being the only game in town, but that advantage won't last long. Therefore, there's nothing to be gained by being the company that takes a chance on Linux. Sad but true.
I guess what I'm trying to say is that most of the discussion on Linux's chances of success revolve around its worthiness as a platform, but a good platform isn't enough. There has to be a strategy to attract gaming business, and Linux doesn't really have one that works. Steam's support is nice, but in the long run it just isn't enough given the risks that an EA or iD would have to take as things are.