I’ve run into that problem as well. Once I center the chuck and the gantry, I place tons of magnets on the Y-rails as well as on the bed, at the forward edge of the rotary. This provides for an easy Y-zero mechanical reference.
If something requires an X-reset to zero, pushing the gantry clear of the rotary prevents collision. Pulling the gantry back into place is facilitated by the magnets blocking the rails.
Reading your post caused another thought to pop into my alleged mind. If your laser has hall effect limit switches, you can trick them into a new zero reference. I usually do that with the Y-limit switch, the one that expects the gantry to travel to its limit, but spins the chuck really fast, then really slow until it times out.
Tap the limit switch on the gantry with a metal object, to illuminate the switch, then tap it a second time to register the new location. It’s not particularly precise, but you’ll get a new zero for the session, preventing the nozzle from colliding. This will cause the opposite defined limit to be incorrect, but you should not be traveling in that direction to the extent that it will collide with the structure.
The new limit will remain until you press reset or until the system is powered off.