Don't Angel also own the 175s though - so presumably they would be even happier to get the 175s on-lease than they would be to shift the 158s from TfW to GWR?
The whole system is pretty daft (although I think I heard/read somewhere that John Major's preferance was a return to somethine like the 'big 4', with regional companies owning track, train and stock, so perhaps the mess we've ended up with shouldn't be described as being "Major's model"). For example, where's the sense in EMR and LNER both having double-digit fleets of 5-car bi-mode InterCity stock when the former is planned to see Kettering to Sheffield/Nottingham electrified which will leave only limited services (routes with only 1-2 trains per day) where EMR will require the diesel mode on their 810s. It would make far more sense in my view for a single pool of bi-mode units, of a single design and livery, to be shared between what (under the current system) are currently XC's, LNER's and EMR's off-wire InterCity services.
I agree that Cardiff-Portsmouth needs to be at least 5-cars (throughout, no spliting/joining - unless using 158s so the guard can move passengers into the correct portion before the split) but I'm not sure Turbos are the only alternative. Would Cardiff-Portsmouth be 8 diagrams? According to Wikipedia, GWR only have 18 class 158s so, if running them in pairs, would only have 9 sets (not enough to reliably cover 8 diagrams, and many would only be 4-car). However, there are 27 class 175s - pairing eleven of the 3-car units with the 11 eleven 2-car units would leave five 3-car sets to support other routes. Alternatively, are the 80x cleared for Cardiff-Portsmouth to cover a diagram or two allowing more 158s to be made up to 5-car?