Hi @alexcussigh0,
Check out our RESCRIPt 12S db tutorial. It's just a simple example of what you can do, but it should help you get started...
Although it means re-downloading everything again, and potentially in batches by taxonomic group. But everything will be formatted properly.
The notebook specifically looks for records of 12S reads, but you can perform a separate search that downloads only genome records and then use feature-classifier extract-reads to extract the ampicon region from those genomes. Then you can merge with the other batches of data.
Anyway, I just wanted to provide another option that you can try running in parallel.
-Mike