The inclusion of multiple hypervisors will lead to higher staffing support costs, additional command and control management platforms, and less standardization on hardware platforms. In other words, having more than one perpetuates the sins of IT past. It doesn't simplify business solution delivery; using multiples lends itself to more silos of stuff.
Everything that is discussed regarding multiple price points in the article is possible from a single hypervisor perspective, it just takes some forethought to how service delivery and licensing is managed. If you want a pool of resources that doesn't offer the latest and greatest in DR or HA capabilities, then the end user (or department) shouldn't get charged for it. Doesn't mean you have to change hypervisors, just change what add-ons are purchased for that pool of resources. By doing so, you stay standardized across the datacenter and regional offices without the need to support an ever increasing ball of complexity.
Virtualization done right is hard enough on it's own. Organizations should avoid falling into the trap of defining their virtualization tools as tactical solutions to localized problems. It's much larger than that. Virtualization done well is a strategic position that maximizes standardization, ROI, and time to market.