[darryl]

@ratinox. I'm impressed that you have gone to so much trouble. But the data to make a calculation such as you have attempted is simply not available to us. Your methodology is flawed and the assumptions you have used are problematic to say the least. I'm not sufficiently interested in this to spend much time on it, but make the following comments:

You estimate Amazon's share of revenue from the sale of ebooks is about $500M annually, based on ebooks sales of 1M per day at an average price you estimated at $4.50. Your starting point, 1 million ebooks per day, is presumably taken from the February Author Earnings Repot which you link to. However, that report also contains a far more suitable basis for the calculation you attempted. That is, ebook sales revenue per day of $5,755,000. Adopting this as the basis for a similar calculation results in Amazon's share being in round figures $700M per annum.

Your estimate of Amazon's electricity usage based on rated capacity? I will simply say that it is very crude and leaves a lot to be desired. For example, are all operating at full capacity? Assuming thay all operate at full capacity, what is the mix of 25 and 30 capacity Data Centres?

Your next step is even more dicey. You concede the report is somewhat outdated. However, what on earth makes you think you can estimate server usage based on an indefensible split half way between percentage of Kindle contributions to Gross Revenue and to Net Profit. A figure which you yourself think is too high and reduce to an arbitrary 10%. Amazon's servers run all its infrastructure and businesses, including all retail sales, all marketplace sales and, of course, its cloud computer business, which logically would seem to be far more demanding of server capacity than storage and sale of ebooks. Percentage of Sales, gross or net, is a very poor indication of server capacity used.

And what on earth makes you think that Amazon would be paying $US0.25 per kilowatt hour for its electricity? Particularly if it is using so much of it?

Unfortunately, your estimates are not even vaguely accurate. This does not mean that your position has been totally disproved. Simply that your calculations are of no assistance whatsoever in resolving the issue one way or the other. However, your position is, in my eyes at least, totally counter-intuitive. My position, at least until is is disproved, is that overhead costs ot the type you refer to, whilst they do exist, are negligible so far as they relate to the cost of storing and selling Kindle books.