Home Automation today is an expensive hobby. To play in that field you need to commit $/E100k. It doesn't need to be that way. The cost of hardware, built at $5 and sold at $500, is there to take care of the ecosystem needed to implement the solutions in the field, a.k.a "the installer".
The reason the margins need to be so big is because the installation and maintenance side of the job are so time intensive. You do few jobs per year, so these jobs need to be expensive. This has confined automation to commercial and high end residential. The cost of automation becomes marginal compared to the whole but doesn't scale down.
To scale volumes and profits, one has to save time. Saving time in programming, installing and servicing are ways to scale the activity up.
Some of the self service tools we are developing go a long way towards that goal. You will find that the iPhone UI we are building is going in that direction with easy to use web based interfaces so that even end-users can engage in simple scene programming. Not all clients will adopt that, but more and more will. The more self-served the customer is, the more time is freed up and the price can come down.
While immediate productivity can be achieved through software-only tools, there is also a big margin available for hardware to go up in smarts. Where standards like KNX are implemented with kb of memory from 10 years ago, today's hardware standards speak in MB. A cell phone runs J2ME or Android. Within the decade, these smart actuators could sell for half the price of their memory limited ancestors with 10 times more features. Can you imagine the day where actuators will export their features for the network to automatically register and export.
We have done that in Java, yes we can. (sorry Obama fever).