Cost Assign: O(n * k * d) Update: O(n * d) Overall: O(n * k * d * iter) where: n is the number of documents k is the number of clusters d is the dimension of the embeddings iter is the number of iterations