I started to do that at AWS. You basically cannot get a 64GB memory machine. I can't remember if I actually got one configured but I do remember calculating that a 6 CPU 64GB machine with 250+ disk space would run about $75/day. That would be in addition to a machine running PRIME. So for 3 months of lab you could just spend $6K on hardware and run both APIC-EM and prime on a host (like ESXi).
APIC-EM will run (not sure how well) on 40GB but as I remember the best machine I could get at AWS was 32GB
Good luck and let us know what you find.