Apologies… I don’t mean to derail the thread. Just trying to be helpful. Also, these are just my ‘internet guy’ opinions, so as usual, take them for what they’re worth. I’ve been doing this for a good while though, so hopefully that counts for something.
When you say analog summing I think something like an SSL Sigma and that’s probably what you mean as well. I personally don’t feel that it’s worth the effort and cost. Does it make a difference? Yes, absolutely, but IMO more subtle and gentle. I don’t think it’s as substantive as a lot of people would make it out to be or more realistically what they want it to be. Probably not the big picture change a lot of people want to chase.
It’s really interesting you bring up the 90’s comment though and I think it’s very relevant, but maybe not in the way you’re thinking. IMO, all that character came from noisy or somewhat noisy analog mixers (depending on what you had), analog synths, lower resolution sample rates and DAC’s. That was the character good or bad. We all chased this constantly moving idea of ‘clean’ and ‘high end’ until now that we have it (hardware or software) we’re back to wanting a lot the sound we largely ran from. I was guilty of doing it more than most probably.
Just yesterday I helped mixed down a track for a friend (well established electro artist) which was all in Ableton. It sounded pretty good, but we decided to stem the mix out to hardware. He still has a Yamaha 02r (I had one too in the early 2000’s) which by today’s standards shouldn’t age well at all. Most of today’s gear, on paper anway, has much if not far better specs. We separated it out, also having to go from 24/96 to 16/44 and it sounded great… better than the ITB mix we had. It had more punch, separation and fullness to it. Why? There’s something to be said for less than perfect gear. Those easily distorted converters have their own character and sound to them. That includes all of the extra conversions since we went analog to the O2r where it converted twice, then analog back to the Focusrite where it was converted yet again. People would have you believe this should all sound terrible, yet, it doesn’t.
We wouldn’t have gotten the same result from ‘better’ converters, really clean desk or going all digital to the O2r and back… we actually needed worse gear and more conversions! We got that tone, flavor and punch. Sure, that’s not going to work for everything, but I’m talking in the context of electro and not recording/mixing something like classical guitar or orchestra.
On the flip side, when I go to master it, I don’t necessarily want big changes… I want clean and subtle (for the most part). We already got the vast majority of the flavor in the mix and I’m not trying to add too much more, so it becomes a lot more about last 5-10% and not the first 80-90%.
And in the end, I might decide it sounded better bypassed after all 
