Hmm, I believe there're many open-source self-hostable office solutions.
And they don't share a universal format.
And they don't share a universal format.
When the format is proprietary, you get a blob of binary data that you have to reverse engineer through guesswork, made orders of magnitude more difficult if the original service isn't even available anymore.
Also PDF is turing complete.
Modern rich content formats are hard. Open formats solves only small part of the problem.
I think the universal interchangable format is static HTML.
Meanwhile you look at actual standard formats like PNG or UTF-8 or, as you say, HTML, and the implementations generally do a lot better because there is a proper spec.