Undersea cables are a pie in the sky; current high-load cables in urban an industrial areas could be made much smaller, simpler, and lossless.
I wonder if transformers, currently huge and expensive, could be made better with this, too; at least the ohmic losses could be removed, and thus a lot of need for cooling, and the fire hazards.
Not really, when the sun is up over the Pacific ocean, there's not that much sun over land. Maybe a global grid happens anyway, but cabling losses aren't the only source of cost, so I'd put my money on more localized improvements.
Better interconnection between and within local grids (maybe a viable Tres Amigas interconnection, but even just better connections between sections of the major grids would help with grid management. Improvements in motors, MRIs, magnetic bearings, transformers, etc.
Superconducting magnets become cheap and widely available which allows for maglev trains at massive scale. Costs for the LHC and similar experiments would drop dramatically. MRIs would only require air conditioning, if that; Modern cell phones are sufficient to compute tomography. Magnetic confinement fusion also becomes cheaper and easier. Electric cars could use superconducting motor magnets allowing for even greater power to weight ratios and efficiency.
Just a few things off the top of my non-mechanical-engineer head.