The sucky thing is, it could be a modeling issue, but the answer would have to be a model that agrees with all observations. For example, it could be that spacetime can get permanently warped such that gravity-like effects remain, but then how would a model represent that? If the model represents it as a field that is held in effect by some localized particle, then that 'something' might still be called "matter" even though it could be nothing more than an artifact of that particular model.

For a similar happenstance with current models, see the "graviton". If spacetime 'changes' due to the presence of matter (at least, insofar as locality and position itself is real) and nothing more, there might not actually be a graviton to discover, yet that's what the models demand to become closer to observed reality.