It's not a case of "if hardware fails" it's a case of "when hardware fails".
Restoring a complex server set up to new hardware from scratch is a time consuming and expensive business.
Most businesses cannot afford such an unplanned outage to mission critical IT services.
Unplanned downtime can kill a business.
Maintaining an exact replica of existing server hardware makes no financial sense,especially when faster hardware could be used for the same price at the point of failure.
Virtualization runs your servers as hardware independent processes