Let A(m) be an m x m principal submatrix of an infinite-dimensional matrix A. We give a simple formula which expresses A(m+1)-1 in terms of A(m)-1, and based on this formula, an algorithm which computes the inverses of A(m) for m = 1, 2, 3, ..., n using only 2n3 - 2n2 + n arithmetic operations. This is an improvement over the naive method of computing the inverses separately which would require SIGMA(m = 1)n m3 = O(n4) arithmetic operations.