There are two directions I see coming, one you probably know about, the other you do not.
The first is PC based sources. The idea of having an audio appliance will come and go, with perhaps some hold outs, but most vendors going to a pure PC based solution. Sorry Apple, you'll probably be left out because of your insistence of all hardware coming from you. Vendors will be looking for PC's because they already read all the software. DVD/Blue-Ray/Internet Radio, and anything else that happens there will be cheap readers for on the PC. Now, some will take the PC accessory route, like USB DAC's, but more sophisticated, either by offering PC cards or perhaps by using custom outboard processors like E-MU uses.
The second, not so obvious trend will be hyper modularization. Companies with the know how and money to spend will realize that the only way to keep delivering the best solutions, and the most flexible solutions will be tiny modules that can be plugged into a virtual network, all of which can be PC configured and driven.
Imagine if you will a rack mountable chassis that accepts 1 to 12 cards, each of which can be a processing card or an output card. That these cards can handle audio or video. That you control how things are routed via a PC based configuration program that allows you drag and drop these cards into your own custom audio / video processing chain. Further, because the system is so modular, and the PC is just a commodity item you end up paying no more than say a couple of hundred dollars more than a mid-range Onkyo or Pioneer processor.
Also imagine that multiple chassis are linked via a simple network cable, and that output cards are all optically isolated from the network, and master clocks and FIFO buffers are all near the target outputs.
After this there may be another open vs. closed platforms battle. Some companies will have entirely closed systems, and others will publish their spec's out to the world to design accessories for. Wouldn't it be great if you could say, get your DAC's from Benchmark, but then buy whatever room EQ you chose? Or have just one media storage appliance in your entire house, and then pick and choose which rooms get what type of playback? Does your half bathroom really need 9.1 channels of 24/192 dCS ring DAC's to play the morning news while you browse yet another lame ass review for magic stones to put on top of your amplifiers? Those companies who have the most confidence in their products will take the open source approach. Those who feel they have a lot to loose if people can pick and choose will go closed. They'll come up with bogus stories about tight integration and better user experience, blah blah blah.
What made me think of this? Perhaps it comes from my obsessively watching for news from Theta, and their unwillingness to start a new platform paradigm instead of trying to upgrade the Casablanca. While they seem to be making noises about releasing an HDMI card in our lifetimes, what I see absent that everyone is going to ask about next is the room correction software. Mind you, I'm an absorber/diffuser sort of guy but the latest generation of acoustics processing software looks really sweet to me. What good is it, people will ask, for an aging platform to add all the latest HDMI audio formats if they can't also integrate it with the latest generation digital processors? When the Casablanca can't help you get room eq, internet radio, or download the latest 24/96 track is it really anything more than a rack of DAC's already?
I don't know what Theta, Meridian, dCS or Halcro will do next, but I do think that when the economy perks back up the high end will once again reconsider the home theater market and be looking for some completely new completely outrageous product platforms, and I believe that those who don't go into making the PC a complete home theater platform will look towards networking and hyper modularity as the next logical places to be. I also wonder how long it will take for smart stereophiles who also happen to be industrial and electrical engineers to start their own open source home theater platforms. Also, my guess is that some smart video only companies are probably already thinking this way and we may see entries from them competing for the real high end space sooner rather than later.