Ultimate is a niche use case - it’s for if you need to do sessions that are over 128 tracks, need over 32 channels of I/O, use 96k or higher sample rates, mix in 5.1/Ambisonic, etc. But if you do end up needing to pony up for Ultimate, it can run natively without needing any special hardware.
Digilink is niche too - Pro Tools hardware is kind of like an Apollo, you offload DSP to dedicated hardware for low-latency tracking through plugins. It’s amazing if you’re T-pain and it’s 2007 and you need to track through Auto-tune. But its outdated and locks you into an insanely expensive proprietary ecosystem. So skip it and just go Thunderbolt.
For another hardware option - for multi-room studios Dante is the shit. It’s Audio-over-Ethernet, so you just plug all of your devices into a router and send anything anywhere with the click of a mouse. You can freely send I/O between rooms, mix and match all kinds of hardware without worrying about clock sync, etc. We’re using it on a studio build right now and I’m just in awe of the capabilities. But it’s a lot more complicated than getting a couple Apollos (what I personally use).
I know Pro Tools is annoying/intimidating but it’s so much more powerful than any other program. Any audio-engineering-oriented task is way easier than in Ableton, like tracking a band or layering and comping vocals. I used to slave Ableton to Pro Tools, with a Push controlling it and Ableton tracks bused to their own tracks in PT - DM me if you need any help getting your head around how it works.
For example, you could jam in Session mode along with other singers/instrumentalists and have your individual Ableton tracks recorded as audio in Pro Tools alongside theirs. Then you can send that session out for other people to overdub/mix without having to print out a bunch of stems, and still be able to change the Ableton arrangement at any point in the process! (And personally I really enjoy working with printed audio because the urge to tweak dissipates.)