Yes - our users run queries across hundreds of AWS accounts and regions. TBH, we never expected it to scale that far when we started, but it has. Steampipe parallelizes the API calls out and performs appropriate backoff in throttling to gather the results.

Steampipe does the API calls needed for the column data requested [1]. So "select instance_id from aws_ec2_instance" is much faster than "select id, tags from aws_ec2_instance" (which does a tags API call per row). But, it will still work better than you expect, and Steampipe caches results in memory to make subsequent queries instant.

1 - https://steampipe.io/blog/selective-select