Hey Ted. That means a visual setup, right?
I think the GoTo should be OK with the 3D printed rings system flex. Because the latter would be less than the error caused by the optical axis of the Newt not coinciding with its OTA axis slightly. Thus, after the roll angle adjustment you would either want to re-align the GoTo anyway or just tolerate the error in the AFOV and adjust for it from the remote. Add to that the GoTo error caused by the not ideal and random orthogonality (dependent on the roll angle of the collimation axis)...
Thus, I would suggest 3D-printing such a ring system in some rigid enough filament (CarbonFiber PETG/ABS or better). Making it a bit beefier than aluminum mass.prod. one should be enough to make it totally reliable. I would add the 3-rd fixed on the OTA stopper ring to prevent the OTA from sliding down under the gravity when adjusting the roll.
I think I can help with the 3D design drafts, though not with the printing, as my BambuLab has only a 10x10" footprint (there is a multi-segmented design possible, but fewer pieces to join is always more reliable).
I would estimate the material cost of printing such a system for a 10" scope at around $20 + metal nuts/bolts of your choice for joints.
Upd: Just saw folks' concerns about the 3DP parts reliability. That's a misconception from the early 3DP days (I've been printing since 2014). I saw ABS parts printed for car and motorcycle frames taking hundreds of pounds dynamic forces. They reported to hold even better than injection molded OEM parts over years of (ab)use. It all depends on the proper design, material, and printing quality assessment. Then a bit of personal observance (if you see a crack starting, just analyze, redesign, and replace the next day).