No? You made it sound like the only way it could go that fast is because you used a proprietary AWS extension. The fact is the original code had to be grossly inefficient, probably doing something like individual inserts outside of a transaction, to even be that slow.
Well it wasn’t going to be that fast outside of S3 and please don’t give the tired cliche boogeyman of “cloud lockin” like using that one extension is going to make a difference