Money could be better spent on improving delay reductions using advanced sequencing from gate to runway by reducing holdups for aircraft moving under their own power.
Exactly.
A-CDM is helping in this respect and the eventual goal must be for all departures to have a minimum taxi time straight to the runway for takeoff. If all the needs, positions, speeds, etc. of aircraft both on the ground and in the air are known, then you can have real-time optimisation. Yes, there will be tech problems, go-arounds, RTSs, stand congestion and the like but clever software should be able to adapt to that.
If an airport is planned to run at X movements per hour, there’s no insurmountable reason why the majority of those movements should suffer any delay in taxiing. It just needs a holistic approach where all operations can be varied for the common good.