I think this is just over-complicating things. My only question was does anyone know how much a P90D would theoretically benefit in performance by not being restricted to one gear ratio. The answer is no, apparently.
The available data suggested there would be virtually no benefit, only a slight advantage at higher speeds - and that assumes your notional CVT is as efficient as the current gears in transferring power.
From 0-30 the current car is traction limited. No amount of additional power can make it move faster unless it is delivered by something besides the tires. The only way to speed it up would be sticker tires or traction compounds on the pavement - or rockets.
Elon's statement said that the P90DL extends this up to close to 60 mph - meaning that no transmission will make the car faster up to that point.
Above 60 there's a constant power range, which a CVT also won't help.
Somewhere in the 80-100 mph range, though, the car starts losing power due to back EMF effects. This is the one place your notional perfect CVT could deliver more power - by slowing the motors down, it could reduce the back EMF issues and allow the constant power region to extend up to the maximum speed the car can reach.
In real life, however, a CVT able to handle the rear motor's torque has never been built to my knowledge. If it did exist, it would be a lot heavier than the existing gears, and lose more power (good CVTs are 90% efficient - the current fixed gearing should be ~98%.)
That means the car would be slower in normal speed ranges and need more energy per mile, but might have an advantage at >100 mph speeds.
OTOH, they could get a very similar result without the extra weight by gearing the motors taller - except it would be more efficient, requiring less energy per mile.
A car biased for high speeds might justify a two speed transmission, though that means more weight and cost and complexity, more things to fail.
Walter