What I mean is that the gap between someone who has studied the new opening vs someone who hasn't studied would be huge so players have a much stronger incentive to prepare. And with an engine the new theory comes out immediately, it's not like you're waiting for humans to develop it.

Imagine a match in classical chess between a 2600-rated plyaer who has spent time preparing their classical openings and a 2800 rated player who hasn't prepared at all - the 2800 player will still have a large edge. Now imagine the same scenario for a 960 game where the lower-rated player has spent 4 days evaluating the opening with an engine and the high-rated player hasn't - in this scenario the advantage from the engine prep is much bigger. The mix of novelty + opportunity to prepare is such that from a game theory perspective the whole thing becomes prepare-or-lose if their overall chess strength is reasonably close otherwise.