Variance and Covariance Twist

When applying variance function over rows axis np.var(matrix_A, axis=1), it returns this array [1.2, 6.16, 6.8]
While this does not equal the main diagonal in covariance output as indicated in minute 1:26
I think it is related to the numerator of the formula to be either N or N-1
so why is it happening this way or am I getting thing wrong??