The science was worked out decades ago -- take second semester General Chemistry and you'll know more about this device than what you'll find in the science "journalism" stories on this topic. There appears to be no *really* new technology involved (membrane cells have been around for a while), though you won't find the details in this drive-by reportage. It *is* an incremental improvement over other storage batteries, such as lead-acid, but revolutionary -- no, not without a fairly generous definition of that term. It is a business opportunity, not a world-changer.
Because it uses solutions rather than pure liquids or solids, it is much too bulky and heavy for portable uses. Also, vanadium is much higher atomic weight than lithium, so that adds to the weight. The "non-toxic electrolyte" is apparently sulfuric acid
-- which doesn't stop proponents from dismissing lithium (once an ingredient in 7-Up) in competing technologies as "toxic". Oh, and both redox reactions are one-electron processes -- a two-electron process would halve the amount of solute required, all other things being equal (an important qualifier).
"two solutions, one positive, one negative ..." Barf. OK, not everyone stuck through the second semester, but that's still horribly inaccurate descriptive writing. In fact, one side of the cell uses the V(V)/V(IV) couple, the other V(III)/V(II). The advantage of this setup is that both reactants and products in both sides of the cell are in solution -- no solids, which leads to complications which limit the lifetime of lead-acid, lithium, NiCd, etc. cells. If the two solutions don't mix, there's no need to use the same element on both sides -- I don't know if there's a compelling reason for it in this device, but there's a good chance these two couples aren't the only ones that work well. After all, the Periodic Table has a little something in it for everybody.
It's worth noting that these cells have been in commercial use since 2000, and they haven't shaken the ground under anyone's feet, as far as I can tell.