Continuous Delivery and Infrastructure as Code is mainstream, isn´t it? Everybody pretends to do it. But why do we not see tremedously improved delivery speed in larger comapnies then? In this talk, we shed a light on the reality and the not so nice aspects of it like:- mission critcality and the stability of pipelines- the "zero to hero" principle or how much should I automate?- Pipelines also need architecture- why cloud?Apart from many fresh ideas to improve the situations given, we also look at real-world pipelines and the run of some infrastructure code.