If you make the cars smaller (call the scale s), they have a smaller frontal area by s^2. They can also go slower and look the same on camera, again by a factor of s. Since energy consumption scales as speed^3 * frontal area, you get a remarkable
power ∝ s^5
So if you make half-sized cars you can use 1/32 of the energy. And you can have an identical visual spectacle as long as you scale the camera positions, reduce the tire grip by s^2, and find some 3' tall drivers.