Ofcourse hydrogen is less energy efficient than batteries, generating energy as electricity to power electrolysis to make hydrogen, and then putting that hydrogen in a fuel cell to convert back to electricity to produce kinetic energy is always going to be less efficient than storing the electrical energy directly in a battery, but I'd say this isn't such a problem really.
Why?
Well, in the sort of ideal future grid we should be working towards there will be ample energy to spare. Nuclear fission power can't be throttled back easily, it runs at close to full power all the time, (fusion power would perhaps be the same once available?) so there would be excess energy in the grid much of the time. Energy efficiency is not the top priority, sure it is nice to have, but the future we should be planning for is one of cheap plentiful energy, exactly as the original hopes for atomic power suggested. When you've got energy to spare, the chemical nature of H2 fuel, and its energy density makes it superior to batteries (not to mention speed of refuelling as vs time to charge up batteries) even if you lose a large proportion of energy in the conversion steps. Futures planning for electric cars optimised for a grid too feeble to supply the true needs of a place are completely missing the point, the only type of grid which can usefully supply an industrialised civilisation is one which is producing the sort of levels of excess power that energy efficiency is no longer the top concern.