The whole point of end-to-end encryption is that you don’t have to trust their server: it cannot read your messages. Then for metadata, the question is about what metadata they are receiving at all (if they don’t receive it, then you don’t care if the server is proprietary) and what they do with it (e.g. for the private contact discovery, the idea is that you can verify that the code receiving your contact list is doing what it should (it’s open source), and you can verify that this code is the one running in the secure enclave.
You don’t need the whole server to be open source: only the important parts.
The whole point of end-to-end encryption is that you don’t have to trust their server: it cannot read your messages. Then for metadata, the question is about what metadata they are receiving at all (if they don’t receive it, then you don’t care if the server is proprietary) and what they do with it (e.g. for the private contact discovery, the idea is that you can verify that the code receiving your contact list is doing what it should (it’s open source), and you can verify that this code is the one running in the secure enclave.
You don’t need the whole server to be open source: only the important parts.