[buchanmilne]

> Amazon is AWS's biggest customer, but I've never been told to compress things with zstd while working at AWS.

You may have switched transparently to using it (e.g. for your application log files), via internal Amazon tooling (that many AWS services use).

However, this tweet wasn't about AWS customer's data, but AWS's own data.

> Maybe data in S3 is compressed? I don't know the internals of S3.

Almost every other AWS service uses S3 for storing something.

Think of EBS snapshots, DynamoDB backups, RDS backups.

And that's not even the service-internal data.