I do agree meatgrinder scenarios are bad for creating backstory, but that's because the entire concept is oxymoronic. Backstory for a game is is story of events that happened before a game starts. It exist to explain the initial position and to hint at follow-up positions.
If you start up a scenario not knowing that the end result will be, that's not backstory, that's just playing the game. To say you're "creating backstory" is just a fancy way of saying you're looking to build future events on results of game play, same as any other time. The scenario itself needs its own backstory and it's turtles all the way down until you just accept some initial position as the premise without further questions.
The actual purpose of "playing your backstory" is to add in detail to a framework that's already known. F.ex. you know your character is last survivor of a village destroyed by goblins, your choosing to play that event to show who the goblins killed, how & in what order. This doesn't apply to meatgrinder scenarios where it might not be known beforehand which characters come out alive.
All of this is completely separate from the question of "do meatgrinder scenarios serve as a bonding experience?" Almost any shared experience can work as a bonding experience and that's why you'd want to play out dramatic (or traumatic) shared character events, because it isn't actually a shared experience for the players before they play it out (duh). Expecting a perfect success rate is daft, for the most part you're hoping to beat a coin flip. (As a corollary, you have to try a method more than once before passing final judgement on it. This very basic principle seems lost on many people.) Creating unit cohesion is even more specific and I laught at the assumption that it is necessary or even desired end result of the method. Abandon performative group play and embrace individual characters trying to kill each other today.
*Ahem* What I mean is, meatgrinder scenarios & other trial-by-fire deals are good for testing
if players and their characters are capable of cohesion. Actually building cohesion takes considerable time, a safer environment and some trial-and-error. In plainer terms, the actual bonding more likely happens in the group therapy or rebuilding session after the horrible disaster all the characters underwent. If you want an actual idea for how to make people function well as a small group, forget roleplaying games and get your hands on a Scout Master's guide or something.