It’s a multi stage transition, and we are nearing the end of phase 1, rebuild in AWS using equivalent EC2 instances. This is not cheap by any means, but it gets us out of VMware, and also gets us away from our abysmal capacity management problems. Being able to keep up with business growth on-demand is worth the extra cost over the medium term.
Phase 2 is rearchitecting and consolidating our major overlapping LOB applications into fewer, more cloud-native designs. This should reap cost savings eventually.