The strange situation is that the 'input' energy source would in fact be the same grid as the grid that power is being supplied to. Theres a crazy pricing scheme which makes this a viable proposition.
My first impression is that one customer's power is going to be sold back into the same grid by a second customer. The "crazy pricing scheme" might simply be that a next-door-neighbor is a high volume consumer with a better per-kilowatt negotiated price. The coupling between the two might be as simple as a heavy extension cord dug under a fence.
Did I guess correctly? What do I win?
If this scheme is some variation of the above, you can be pretty confident the power company attorneys anticipated this and included penalties in their customer contracts for trying it. Please keep this thread updated as events unfold!
As a purely academic question, this raises a topic I've long wondered about: How do multiple grid power sources (dams, nukes, coal, NatGas, whatever) "push" power to the shared grid. At first glance it would seem that any individual source would need to increase their voltage to insure they are a source and not a sink, but that would quickly lead to a game of oneupsmanship with little increments in voltage causing each to further bump their voltage, ad infinitum. A bit of research revealed that in the interesting world of multiphase AC,
frequency is used to measure and control this behavior. If your frequency is lagging, you are a net sink. So the goal of power plant control loops is to precisely match the frequency (and, of course, phase) of the grid. Doing so insures you are not sinking current.