I suggest instead of the awk one-liner you look at gz-sort, it is a small linux program that sorts gzip-compressed files on disk while using a very small memory buffer, as low as 4 megabytes.

You sort the file using

The -u switch removes duplicate lines from the sorted output, and you can increase the buffer size to give it a larger buffer for transporting stuff, but this isn't necessary. I used -S 1G to give it a 1 gigabyte buffer and it took around 7 hours to complete so not much shorter than the advertised completion time, 9 or 10 hours. So this program will run well in your VM, the RAM factor isn't important.

You need to compile it yourself using make but it has minimal dependencies, only zlib and GNU headers.

I used it to find the smallest address in the dump using

This prints 1111111111111111111114oLvT2. This address was used 55405 times (!)

Here are some the other smallest addresses:

---

Maybe you can also make a list of addresses sorted by balance, now that you have an efficient way to deduplicate them.