I suspect they've spent much longer looking at 2d shapes than 3d ones (because they're easier to imagine and reason about, and there are fewer variations in how "aperiodic monotile" can be defined in 2d).

Also, the tricky bit for this kind of shape (similarly to Penrose's tiles) isn't finding the shape, but finding the aperiodic tiling and adding gadgets that rule out any periodic tilings (and proving this). The actual aperiodic monotile shown looks like it has 201 faces (although that could trivially be reduced by quite a bit).