Dropbox Storage is a MUCH easier problem to solve than S3. They don't have the mix of high perf vs cold backup storage, so design is much, much simpler.
I'm too lazy to find the slide deck from AWS CTO where he explained it, the bigger the total scale the easier it was to support bursting workloads by spreading chunks of data across the whole fleet of disks.
The bigger the scale the easier it is to support high perf on HDDs.