Feels like Maya is still the default tool for previs, layout, rigging and animation. There's so much muscle memory in those roles, and so much reliance on legacy scripts/toolkits in animation and rigging, they'll be hard to shift.
Anecdotal evidence is that Blender's anim and rigging tools are still not quite up to the task for heavy production assets. The handful of riggers I know will occasionally compare mGear+Animbot to Blender, quickly scurry back to Maya. You see more and more Blender rigging tests pop up on linkedin, but I wager no-one is willing to bet a commercial or a show on Blender rigging right now. Maybe improvements in 4.1/4.2 will change that.
Doesn't feel like any other alternatives are ready either. Houdini and Apex are still a bit too early, I don't know of anyone who's adopted Cascadeur or Rhumba. XSI is properly dead dead (took ages!), I don't think there's any other serious options around.
You'd hope Autodesk are paying attention, but I fear what's happening is the unfortunate expected result of Alias -> Alias Wavefront -> Discreet Logic -> Autodesk company swallowing/rebranding. The Maya division is so small compared to the rest of the archivs/engineering sectors, any shouting for dev or leadership is getting ignored, Maya is unfortunately stagnating.
I don't bear Maya any bad vibes, I *want* them to succeed and blow us all away with innovations. Competition is good! But all I've heard is complaints from users saying 2025 is a really underwhelming release, or making huge bitchy rants on linkedin about all the bugs that haven't been fixed since v2.