Interbank transfers are very cheap. A lot of central banks have invested heavily in payments systems too (the BoE is one, Faster is very good). But at certain other points, this technology could be very valuable. Two examples that occur to me immediately are pensions and fund administration.
Obviously, Ethereum goes beyond that with a more generalized VM. From what I have seen, this is a ludicrously inefficient way to perform computations. If that is wrong, then maybe there is more. But just blockchain alone will have a huge impact (if you look at back/middle-offices in banks/fund managers/pensions/insurance, I remember working in this area and going to one of the largest pension admins in my country...I remember vividly opening the door to the pig pen, and it was just a sea of people, 95% just responding to very basic customer queries/doing incredibly basic computation work and costing huge amounts of money...because, of course, there is no cost pressure from "customers"...a parallel that more people understand is US healthcare admin, huge inefficiencies, pension admin is like that).
We are not even at day one with this trend. I am sure people will find other applications after, but applying blockchain/distributed storage/distributed computing/whatever to back/middle office in finance will be huge. It also isn't really a question about whether it is going to happen...it will.