IVs should never be reused. In that scheme the same plaintext results in the same MAC which means IV reuse.

I don't remember my original reasons for that sentence. I have softened the wording.

It's been awhile since I researched this answer but if I remember correctly there was a premium on the networking hardware which could support this TPM facilitated authentication procedure.

If authenticity is desired there is mutual TLS, noise KK/KX or some other method of authenticating as well as encrypting a message. I'll admit I read "encrypt" rather than sign and assumed the goal was secrecy rather than authenticity despite other uses of the word authenticat(e/ion).

My web credentials are a little out of date but I thought adding the username to the hashing operation was a mitigation for accidental common password/salt.

A hardcoded salt is worse than useless, it's actively dangerous.

Boy this comment section is wild. Don't encrypt firmware updates? I want to know what corporation allows their security professionals to run around with that opinion. Make my job 100x easier.