From RAX earnings call
We’re actually optimistic around our ability to decrease our cost of compute on whatever compute service we want to look at or network service we want to look at by taking down capacity at just the right time and enhancing the design and performance of these facilities as well.
How much is this cost savings overstated … optimize your MW consumption? But space is space, and RAX has to size compute to peak loads, right? Also, how is this cost not passed onto the customer (i.e. doesn’t the competition data center ops use the same best practices?