You know it's possible to store and serve encrypted data, right? It's not a one time use will-self-destruct-in-30-seconds deal. The data is still decryptable after it's been sent.
Except not really with proper E2EE. Go and join a new Signal or WhatsApp group. You'll notice they're empty. As you were not around for a key exchange when the messages were initially sent.
It's possible to implement shared history systems where newly invited members of a group can request access to the history of a group, while preserving E2EE security.
Bringing up the ephemerality of A/V chats is not about security, it's about the user experience.
> People expect to join servers and have the history available to them to search. E2EE means that history is not available
It's more acceptable to Discord users for video and audio chats to be E2EE because Discord has always offered them as an ephemeral experience: unlike text chats, they never offered audio/video chat history as a feature for users.