The NULL-result of the Michelson-Morley-experiment (MMX) has a theoretical, let’s say geometrical, explanation. You might see it below: If light goes up with 90 degree and comes from that point to the beam splitter.
If the MMX executed here on earth, using visible light with wave length, the ratio of motion (of the earth) w.r.t the light is one into ten thousand: 0.0001. The difference of the x-position that needs to be focused will be two thousands of a nanometer. If you turn the table of the experiment (like Albert Michelson did), the light coming along Fermat's path will vanish because the path length is different by angle.
Hence, the NULL-result of MMX was a very trivial consequence of the setup: The setup works as selector that removes any non-constant path (in terms of length) from the visible fringe-shift result.
Interesting enough, Albert Michelson detected in 1887 the Sagnac-effect, without investigating it.