The default airbus is buggy and low fidelity, FBW is a significant improvement if you are on PC.
In the real world the magenta speed target on the PFD will not necessarily always be exactly the same as Vapp in the approach page of the MCDU (even when in final landing configuration) due to ground speed mini, but it will never be less than it as here.
The picture you’ve shown here is a result of a really fairly poorly modelled default aircraft. In so far as we can tell the 121kts is actually (almost) correct as it’s sitting (almost) 5kts above VLS on the strip which is what you would expect to normally see in light headwinds. Assuming FAC calculated VLS works correctly in the default aircraft (big assumption!) then ignoring any wind corrections VLS+5kts on the speed strip would be the correct speed to be flying.
Note that while the managed speed target would never be less than Vapp in the MCDU, on older airbii you can actually get indicated VLS between the strip and the MCDU to disagree if the entered weight is wrong as the FMGS calculates it assuming landing weight in cruise and using current gross weight after activation of approach phase. Conversely VLS on the speed strip on those older aircraft is calculated by the FACs using aerodynamic data (newer still
Use the FACS but calculate using weight data for consistency). The change means you wouldn’t expect to see a disagreement on a neo, but there is still fairly often a difference between VLS on the speed strip and in the MCDU on the older aircraft.
Vapp = VLS + 1/3 of the headwind components (limited to VLS + 5 as a minimum and VLS + 15 as a maximum).
For landing, VLS = 1.23x VS1G of the selected landing configuration (full or 3). VLS conf full = Vref and would be your base for any approach speed modifications due to failures.
The magenta managed speed target on the PFD should largely be the same as computed Vapp, but as mentioned you may see it higher at times due to the effect of ground speed mini.