Aside from the much-touted uses in cryptography, probably the biggest use will be as universal quantum simulators. A lot of quantum simulations are infeasible beyond trivial interactions, because the classical memory and number of computations required to capture and progress the state of the quantum system grows exponentially with each interacting element in the system. Simplifying exponentially increasing memory/computation time requirements to linearly increasing qubit/simulation time requirements will make a lot of chemistry and material science way easier.