Are you sure this sketch is accurate?
I think F1 teams actually used similar shapes but with the radius on the other edge, the one at the end of the flat horizontal portion. However, purpose was not to create a cavity, but to move the diffuser starting point as much forward as possible.
Indeed, the regulation defined the floor limits, where it had to be entirely flat and horizontal (in car referential), with the possibility to curve the periphery upward with a maximum radius of 50 mm (not fully sure of this dimension, it might have been a bit different). So, instead of just applying a R50 tangentially to the diffuser slope (blue sketch), they would apply a full quarter circle as per red sketch. This would involve this small cavity between the radius and the diffuser, which has no purpose in itself, but if it had any downside, it was probably advantageously balanced by the gain coming with the diffuser starting more forward.
With Formula 1 aerodynamics being so over-regulated, one must always wonder if a shape is there because it has an actual aerodynamic purpose on its own, or if it is just a consequence / side-effect of engineers trying to achieve something else while reading the rules with a different angle. It highlights how pointless it is trying to understand F1 aerodynamics looking at the bodywork without a good knowledge of the regulation.
but if i can sort the cooling issues out it brings me to the back of the pack of those with 100% power.
