I hope that I can do your question justice. Short answer "YES". What you are proposing sounds good, but I do not recommend trying it. I will try to simplify and will be assuming a delivery system typical of the US.

During a lightning strike in the vicinity of any building there are induced on all lines.

By grounding at a single entrance point for the entire site system it seems that the lines at each building would rise and fall together (similar to a bird on a wire effect). As the system is over a large area transients through the ground versus through the wires will differ and there exists the distinct possibility of large qradients in remote buildings with respect to earth and arcing through whatever (destroying equipment) or whomever (destroying lives) is convient will occur.

This effect is much worse if a remote building takes a direct hit.

By grounding the equipment grounding conductor (EGC) at each building (or the nuetral if no EGC exists and no other grounding paths exist between the buildings) it allows for a more direct path for induced currents to get to earth and for localized gradients to be minimized. Each building may rise and fall on its own smaller scale reducing the potential for damage or personnel hazard.

This scenario works much better if each building has a Surge Arrester installed which limits the voltage on the other conductors.

Now wait - there are some big spread out buildings that are bigger than many multibuiding sites, they seem to be installed the single point grounding as discussed in your question?

I do not deal with these types of buildings, but I assume that they are designed (either knowingly or by accident) properly protected from the gradients by the use of one or more of the following methods: A continuous uffer (concrete encased electrode) ground installed (the rebar tied together). The building steel as an electrode system and properly grounded in multiple regular locations. Lastly a ground ring system employed around the entire facility. (A properly installed lighting protection system also helps mitigate problems.)

These systems have (except for a lone ground ring system) far more conductive mass than a Grounding Electrode Conductor (GEC) of your proposed system.

The EGC between buildings will have current low on it during a lightning strike. This is not objectionable current but very welcome (unless you have copper communications lines between buildings which I hope you've protected well at both buildings and are not using them for data, fiber optic is only way to properly handle data between buildings). The current flow is natures way of equalizing the gradients.

Others may have more to add or corrections to make.

Shane