It's fairly safe to say that, in computing, speed (in this case, a form of latency), always makes a difference to someone.
In a cloud environment, that latency could be wasted money in paying for an instance not doing work. Sure, it's averaged over that long-running time period, but in a continuous-deployment environment, that might not be all that long.
It can also have a very significant effect on MTTR in case of an outage, especially if multiple deployments are required.
