My initial thought was: why not fork LibreOffice and spend the extra dev time closing the gap between what it is and what they need?

But after some thought, I feel a cloud collaboration suite makes more sense as big orgs often run on online-first solutions like Sharepoint. So they can tick the essential boxes by being an online collaboration suite, and fill in formatting features later.

Though your points on speed and architecture do make me wonder if Python was their best choice...