I reckon that VoltDB uses compression but I'm not sure about that. Mike Stonebraker, who is the main creator of VoltDB, used to work on Vertica, which uses compression heavily. I heard about a case when somebody loaded a few GB of sequential integers and after loading the data the database size was barely about 1 GB.
VoltDB is designed for maximum throughput & transaction processing performance, so it does not compress data for tables, views, or indexes stored in memory.
We have a blog post on database sizing, or how you can estimate the size of RAM needed for a use case and schema.