In a nutshell, without getting to deep, osmosis is a process of equalizing concentrations on both sides of a semi-permeable membrane.
Without resin, it's obvious that
fiberglass products, mat/roving/cloth have basically zero resistance to the
passage of
water.
We add resin to make a relatively impermeable
hull, and the different types of resins have different degrees of permeability,
epoxy being quite low.
Gelcoat is actually pretty good at keeping the outside water from getting to the fiberglass/resin matrix, but it does nothing to keep the inside water from trying, (by osmotic pressure,) to penetrate the matrix.
The exterior gelcoat surface is thin and weak and has a lower bond strength to the matrix than the individual matrix particles do to themselves.
Interior heat speeds up the process, the heat in the warm wet
bilge wants to transfer to the cold outside water, this accelerates the transfer of molecules thru the
fiberglass matrix until it reached the more impermeable gelcoat.
Result, the gelcoat bond breaks, and a blister of water forms.
So, we grind out all the blisters or do a gelcoat peel and add lots of low permeability/high bond strength epoxies.
Much of this phenomenon can be prevented by making sure that the whole inside of the hull is well coated with a high solids/low permeability coating.
You don't want raw fiberglass under the waterline, and a simple sprayed on thin coat of
paint, while better than nothing, cannot match a good thick coat of a two-part product.
I worked on a 34' sailboat that was hauled into a heated building, the gelcoat was peeled, tarps were taped around the waterline and small heaters set-up under the tarps.
The boat sat there like that for one whole year.
When it left the building and went over the scales it was a few hundred pounds lighter, all in the weight of the water that came out of the hull.