Sort of, quite possibly, if by degrees
This is a speculative, but plausible, answer
First of all, how do qubits interact and states evolve with time?
The description of how individual qubits evolve (i.e. a single qubit gate operation) is given by some Hamiltonian1. Multiple, non-interacting qubits (that are exactly the same) therefore evolve using multiples of that same Hamiltonian. However, as soon as you include some form of interaction, simulating the exact evolution of a large number of interacting qubits quickly becomes intractable (which is exactly why quantum computers should be useful, in theory).
Now, how are quantum processors designed?
Designing the microarchitecture of a processor (such as the connectivity, layout of the chip etc.) is one thing, but engineering sizes of qubits, waveguides etc. requires going on a computer and running detailed simulations a lot of the time. If there was a way to simulate large Hamiltonians for a long time, it's reasonable to assume that this would help improve knowledge of how the qubits interact with each other and the environment, such as in a more detailed version of this and extensions thereof, as well as generally being able to look at how multiple qubits interact at once. This would in turn allow for improved design of the details of chip, which would lead to effects such as reduction of decoherence.
Finally, what are quantum computers good at?
As mentioned above, quantum computers are potentially useful because of what makes them hard to simulate - a quantum computer is a system that quickly becomes intractable it simulate with increasing numbers of qubits. However, something that quantum computers offer a speedup of is... Hamiltonian simulation.
In other words, by iteratively running a quantum Hamiltonian simulation, using a classical computer to (tell the quantum computer to) vary certain parameters, it's not unreasonable to assume that a quantum computer could help for optimising certain aspects of the chip to e.g. reduce decoherence times, improve fidelities etc. In turn offering better simulations and allowing for yet more qubits, which potentially offers better simulations and the continual improvement would (hopefully) begin.
As for whether it does, or if classically simulating a few qubits and extrapolating from this would give a just-as-good design is something that only time will tell.
1 Having said that, in linear optical quantum computers (at least), unitaries are implemented directly using physical components such as beam splitters and phase shifters directly describing the unitary, as opposed to, say, superconducting, where applying microwaves is described in terms of a Hamiltonian, which is then used to give a unitary. OK, all this is maybe simplifying a bit, but that gets the gist across. What does matter is that, in linear optical quantum computers, generating the photons on say, a ring resonator, is described by a Hamiltonian