My team has hundreds of terabytes of compressed data stored in s3 so I sometimes jump up to the limits when I am doing data pulls for analysis.