Im not sure if anyone can answer this question, but I think the answer would be an interesting theoretical benchmark:

if two players are playing HU, and one player (our hero) can see his opponents cards every hand (but opponent is oblivious) what kind of bb/100 would hero be playing at?

Lets say that they are both playing 100bb deep, and villain reloads after each bust. Hero is playing max-exploit, villain is playing a balanced strat.

what about at a six max table, where hero can see all the cards but nobody can see his?

I am basically asking what the ceiling of bb/100 is in HU/6MX format if hero could telepathically read his opponents.

I beleive it was David Salansky who wrote something to the effect of: "the closer you can get to the line you would have played if all the cards were known to you, the better you are playing"

So what is the absolute ceiling of bb/100 if we always played the perfect super-max exploit line?

The winrate in that scenario is totally dependent on the opponent's playing style. The looser he is, the more you win.

Check Pot rippers winrates

I’m not sure it’s really interesting. Against a GTO opponent who doesn’t know you know his cards your winrate would be obcene. Every hand you can take the line that maximizes the money in the pot before you make a bet too large for him to call with his specific combo. The deeper you are, the bigger chance you can do this. It’s free money with almost no variance.

I actually think this is a pretty interesting question. Against a GTO opponent in your example:

1. HERO will be able to value bet perfectly.
2. HERO will be able to use pot odds perfectly.
3. HERO will be able to fold perfectly.
4. HERO will be able to bluff perfectly.

However, against non-GTO opponents, HERO will not be able to do any of those 4 things perfectly but will still be able to profit.

Variance will still be a factor both preflop and postflop and it's something that can't be avoided. The further away your opponent is from an optimal strategy, the more variance that will be experienced.

I don't know how to calculate what a potential winrate can be but we do know for certain that it will be more than 1BB/100 and less than 10000BB/100 due to variance (this is for a HU match).

Even against a non-GTO opponent who doesn't know we can see his cards our win-rate would be pretty obscene as he'll be somewhat predictable. I think the question is more interesting against an opponent who knows we can see his cards, who can minimize the edge we get from seeing them. Our win-rate will still be pretty high.

The bolded is not true. A counterexample would be an opponent who always folds. This is very far from optimal strategy, and the variance of this game is 0.

I see how you got your upper bound, but not the lower bound.

It depends on the average pot-size, but I'd guess it's about 350bb/100 heads up. IIRC, Potripper won at something in the region of 3.5 to 5.5bb per hand. Or maybe it was even higher. He had a VPIP of 100% I think, and a lot of hands saw flops, and he hardly ever folded unless someone flopped/turned the nuts.