I'm not very familiar with the Karl Fischer method, but my guess would be that the H2O would be counted. It may be incorporated into the crystalline structure of a solid, but it is still water - I would think the Fischer method would pull that water out of the crystals. Plus, there are crystal structures with and without incorporated water, but the one that's "anhydrous" won't have any water in the crystalline structure.