How can I optimize the below for loop in Python?

The code is shown below. Is there a way to vectorize the for loop?

import numpy as np

N = 1000
d = 1000
beta = np.random.rand(N, d)
target = np.random.rand(N, d+1)

beta_prime = np.empty_like(beta)
for l in range(0, d):
    beta_prime[:, l] = beta[:, l] + np.sum(target[:, l+1:d+1], axis=1)