One of the main selling points of the cloud is dynamic scaling which necessitates that Amazon have enough servers for some multiple of their customers’ base load. As long as internal Amazon resources are given a lower priority and booted in favor of customers during leak load, the opportunity cost is basically zero leaving just the marginal cost of electricity and hardware maintenance.

MMAesawy gave the correct answer. It makes it clear what Alexa is really costing Amazon in an format that's easy to interpret. This is how most companies handle "internal purchases" from another business unit that also sells that product externally.

Consider it this way: if Alexa doesn't buy that capacity, someone else could, so it's important to capture this opportunity cost.

It's not perfect, however. I've seen examples where the repairs division of a company had to buy parts from its distribution branch at retail price. All well and good, so far. Except they then had to mark-up that part's price in the repair cost. Combine that with the fact distribution would give discounted rates to other repair companies to secure business and what you have is a repair division that's being outbid on price by competitors using the same parts!