Cost curve and influence of complexity

So, I have been thinking about the cost of services and the effect of reaching scale.  When I discuss IT costs and systems with folks, most people are aware of the cost versus use curve and that as the use (size, number of users, etc) increases, you reach a scale of implementation where the cost to add more capacity flattens and each addition units cost is minimal.

As an example take email systems.  From costing spreadsheets we have done on an email system, it takes about 50,000 mailboxes to reach the flat part of the curve where users (mailboxes) cost $10 per year.  The beginning of the curve is steep, so 8,000 users might cost $40 per year per mail box, 12,000 users bring it down to $25 per year, etc.

I hypothesize that as the number of users increases to very large numbers, the cost curve actually inflects upward again.  So, at some point, adding users make the complexity of the system sufficient that it begins to cost more per user.  (see attached graph)

I do not have any proof or cases that can demonstrate this so I would love to know if they exist

Cost-complexity curve.pdf29.33 KB