The heart of your concern is the non-locality, right? The way to relieve your mind is to think of everything locally instead of globally.
We first learn about energy conservation as a global law: The total energy in some (large) region remains constant as a function of time. But there's another, local formulation of energy conservation: The rate of change of the energy density at any given point equals the rate of energy flow into that point. Formally,
$$
{\partial\rho\over\partial t}=-\nabla\cdot{\bf J},
$$
where $\rho$ is energy density and ${\bf J}$ is energy current.
(Technical note: Actually, $\rho$ is part of a 4-vector and ${\bf J}$ is part of a rank-2 tensor, and they're usually called other things rather than $\rho,{\bf J}$.)
Roughly speaking, the second version says that the energy conservation accounting works out correctly point by point in space, which is a stronger statement than global energy conservation. In fact, global energy conservation follows from local energy conservation, but the reverse isn't true. A world in which energy spontaneously disappeared in this room and appeared in Antarctica, with nothing flowing from here to there, would satisfy global but not local energy conservation.
The theory of electricity and magnetism satisfies local energy conservation, not just global energy conservation. That is, when you take into account the energy stored in the electromagnetic fields, you can keep track of the energy density and energy current at each point, and the books balance everywhere. Moreover, the energy density and energy current are well-defined functions of the electric and magnetic fields, so they change smoothly and causally with time.
Once you adopt that view of energy conservation, it becomes obvious that energy is conserved, and there are no non-locality or causality issues, even in situations such as the ones you describe. The energy flows around smoothly and causally, in such a way that the total is conserved.