Foam is mostly air, but the divinycell and other products are made explicitly for this purpose - they're intended for use as composite cores. I'm sure that this must be re-hashing discussions that have occurred many times before, but when you suggest that fiberglass / foam / fiberglass must be "weaker and more prone to failure" than fiberglass / cedar strips / fiberglass, I would think that it depends on what you mean by "weaker."

Given that the SoF people say that a sheet of Dacron coated in urethane is plenty strong enough to keep the water out, I would imagine that a layer of fiberglass over each side of a 1/4" core of 5lb divinycel ought to do just fine at that job, and it would be significantly lighter than the equivalent part made out of wood.

Whether it would be stiff enough to avoid flexing and twisting is another matter - and from what I've read, I assume it would not be. To fix that, you could go with denser or thicker foam, and/or add layers of fiberglass, but at some point you'd have lost any weight advantage versus wood. I'm sure many a boat designer has done this calculation, and the fact that they're still sticking with wood speaks for itself.

That said, I'm not sure whether a boat designer would have had any reason to explore the combo strip/foam approach that I've described. It wouldn't be relevant to a production boat built in a mold, and I don't think it'd work in an S&G either because you're relying on certain characteristics of the plywood. So it'd have to be a strip kayak designer - and they presumably (and for good reason) love the visual appearance of the strips, so why would they bother with a construction technique that doesn't have that?