I was playing around with the Dynomation program I purchased some time ago, designing a theoretical 14.0l 48valve V-12 (based on the Chevy bigblock, with Thunderpower 4valve pushrod heads).
In that program you can make a comparison of up to 4 engines, so i made 2 different intake lengths, and noted down the differences.
The long intake is 3.56 times longer than the short one.
Long intake max torque @4000rpm, short -508nm. (short -286hp)
Long intake max power @6200, short +140hp.
Short intake max torque @5600, long -129nm.
Short intake max power @6800, long -273hp.
The crossover point is @5200rpm, over that the short intake is better.
But both are nearly identical under 2400rpm.
So between 2600rpm and 5000rpm the long intake gives a torque increase of minimum 3,5% up to maximum 37% then down to 6,5%
At 4000rpm the difference is 286hp and 508nm. (In 1995 the BMW M3 3.0l engine had 286hp and 320nm. A BMW 750i (5.4l) from the same year has 490nm. )
I know that the workload would be significant, but I think that, maybe as a later update adding variable intakes, this would give a very real increase in effectivness to the engines.
It’s only a test in theory. The engine only exists inside my head.
It would be fun to try and build it, but it needs a custom casted engine block, custom cylinder heads, custom crank, oil pan, cam shaft, intake manifold, ignition system.
The only parts already on the market are : pistons, valvetrain, connecting rods, cam drive, water pump.
Not just Variable length intakes, but multi-runner intakes, too.
Like the Aussie Ford Falcon, Hyundai Terracan V6, few mitsubishi V6’s, that have two sets of intake paths, a long and short, with throttle butterflies and an actuator.
Like variable cam timing, but for your inlet manifold.
A BMW 750i (5.4l) from the same year has 490nm. 
)
