The difficulty has to increase by two orders of magnitude for these to become totally worthless in terms of power usage versus mining income. I plan to run them for the next couple of years. I'm not really worried. If they lose money, so be it, at least they're not costing me anything to run and they are securing the Bitcoin network.
The per watt efficiency is similar to other manufacturers out there, so as long as they are mining profitably, probably I will be too.
In the meantime you're welcome to disseminate your money to operations who will ship in "September" or "October" or whenever, who will probably use your funds to buy ASICs, destroy the network difficulty with them before shipping them out, and then send a slightly profitable brick to your doorstop (as we're seeing with Avalon and BFL, and as I'm sure we'll see with the Klondike etc manufacturers).
I am a bit confused by your thinking here. You plan to run your USB miners for 'a couple years', then turn around and say that other operations will destroy the difficulty and send something that isn't much more profitable than these things in the first place. For this statement to make sense, the USB miners would have to break even before other ASIC companies ship, which probably isn't going to happen. For the same cost you can get ~15x the hashpower but have to wait, wouldn't that mean that the difficulty would have to rise even more in order to make those options worthless as well?