DOT PRODUCT IN DEPTH DIRECTION FOR THREE-DIMENSIONAL ARRAY WITHOUT FOR SENTENCE

I recently started using python for numerical calculations.
For example,

import numpy as np
A = np.zeros (500, 4, 4)    
B=np.identity(4)

Suppose you have a three-dimensional array A and a matrix B.
Suppose A is a zero matrix, but it is actually nonzero.

for i in range (0,500):
    B = np.dot (A[i, :, :], B)

Towards the depth of the elements of 3D array A,

A[499,:,:].dot(A[498,:,:]) ….dot(A[2,:,:]).dot(A[1,:,:]).dot(A[0,:,:])

in determining the specific dot product such as If you use the for statement like the one above, it will take a long time.
Is there any way to describe faster calculations with the same results?

Supplemental
I intend to repeat the process of calculating the 500*4*4 three-dimensional array A in order dozens of times above.
I want to use the results for fitting as well, so I want to speed up the processing speed as much as possible.
This loop was the most time-consuming part of the process, so the current situation is painful.

It depends on the machine you calculate, but in my environment, the three-dimensional array is 1,2,…,499,500 in the depth direction. The time required to calculate the dot product is approximately 0.4 seconds.
I wanted to reduce my goal to about 0.1 seconds.