It’s not necessarily; you can use a polynomial commitment based on postquantum assumptions, like lattices