At this stage of development these ABP "toys" may be "cost effective" but as these surveillance drones become useful thay are going to cost on the order of a thousand times more than they do now (>$30k/ unit).
The cost of the these aircraft lies not in their propulsion system or airframe but in their software and surveillance systems. Daylight imaging is cheap but night imaging of any useful resolution is astronomical. Reporting the position of the surveillance platform is realtively economical but reporting the position of the target in realtime is economically numbing for the private sector and their are only a limited number of sources for the navigation/flight/communication code. The individuals who developed this sophistocated software over the past 5 years aren't going to donate their work product to the ABP because they're patriotic.
One of the leaders in Micro UAV technology recently reminded me, when I broached the subject of ABP's experiments, that the cost of this technology is only justified by a free spending government because the system saves lives. The ABP will sadly conclude, after their experience, that it is cheaper to acquire, operate and maintain a light, civilian aircraft with a flight crew of two than it is to attempt to operate an autonomous, integrated, night surveillance system utilizing UAVs.
I don't buy it. If you can put the machine up there for $30k. If it lasts just 3 years that $10k a year.
As to software, I am sure the government has the specifications down that they want and can turn to other developers if the current developers try to jack the price up too high. It might mean that it's a few more years before they are being used in mass, but it's still a brilliant idea. And you can be sure the government isn't going to pass on it, because a couple of developers are going to get greedy.
What?
Add an IR imager and thermal cooler for imager, the correct optics for use in the IR range and BINGO - thermal night vision ...
What are you planning on using - a ReFLEX 50 network?