The fact that someone uses AWS for their servers or storage should have no correlation to an individual's Amazon account.
Amazon already has some pretty substantial track and identify capabilities, Thunderbird should be no factor as far as Amazon is concerned.
The question boils down to:
- Exactly what data is Thunderbird saving?
- Can Amazon read it?
- Would Amazon have any interest in reading it?
Amazon is certainly capable of reading AWS content, but their business model would be destroyed if it was a common practice, so outside of a legal directive it's unlikely they have any interest in reading AWS content. Automated hash checks for "Known Bad" may occur.
Apparently Thunderbird defaults to saving a disturbing (to me) amount of meta data. Nothing I've seen implies that they are saving any email, but I have not personally verified that one way or the other. There are controls to turn off sharing Thunderbird meta data, I assume they work but I no longer use Thunderbird since TB78.
Incidentally, many free email services like Gmail, AoL, and others say right in their terms of service that they read your email.