Adhesion seems fine! If you can get the filament to actually extrude! 
Using the Bambu high temp PEI on the A1M and a bit of glue stick. The latest beta of Bambu Studio is also a bit more intelligent about model rotation; it positions the benchie such that the long axis is aligned with the bed movement, which will help adhesion in general.
I’m currently printing this TPU wallet using TPU HF in the X1C. I tried printing it on the A1M – same filament, nozzle size, etc – and immediately got a pile of spaghetti. The X1C on the other hand looks like it will complete perfectly fine (in about 50% of the time compared to the regular TPU profile from the maker). I think I’m done with experimenting with TPU HF on the A1M for now. Maybe I’m missing something really obvious??
Also of note: the X1C also successfully performed flow rate calibration on the filament, setting a value of 0.946 as compared to 1.0 in the default TPU HF profile. Turns out you can’t do flow dynamics calibration with TPU so I had to skip that one.
In the X1C I’m trying out (for the first time – so much for isolating variables) the Lightyear G10 Pro build plate. Not using any adhesive on that one and it’s holding up like a champ. In fact TPU seems to stick to it a little too well. I figured I’d give that a try since it’s an extremely thick build plate and this model has a precariously thin footprint.