Yes Amazon has very low margins, in fact I think usually it isn't even in the black, which is ridiculous for a company its size. Frankly I don't know why they do it, you'd think they should exploit their dominant position a little more and make more profit.
Amazon's been struggling the more they introduce new schemes, products, and services, increasingly becoming a conglomerate and trying to "force the market," pushing funds to stuff like Lab126. They posted a tiny loss in '12, but made ~$274M in '13. They're estimating effectively breaking even this year, and are expected to post weak profits next year. They used to post many hundreds of $millions and even $billions per year. Amazon grows rapidly, but the stock's basically ev- or neutral at best mid-term and long-term almost any way you look at it. I'd argue they have been exploiting their dominant position (or at least trying to), and failing largely because of it.
Basically, I think they've grown too large for their pool and don't have anywhere to go, but they don't really know what to do with their money because they're expected to be multiplying revenue figures like madmen, so they throw it into bad projects. From an investment perspective, it'd probably be good to "go back to their roots," look at their "real" business (distribution/sales platform), and look at ways of squeezing more profits out of their revenue, which'd obviously include accepting BTC (and dumping Prime

).
If Amazon were able to convert just 5% of their fiat sales to BTC, assuming an effective 3% overhead/fees on CC & .5% on BTC, they'd be posting ~$100M less in expenses this year without revenue loss while also ignoring any customer acquisition from the move. -But Amazon acts like a consumer-oriented charity while Bezos sells out of it instead of getting rid of the things generating both revenues and net losses, so with its Profit Retardation philosophy, other companies like Exxon and Shell may actually adopt BTC before Amazon, while also being more important steps for Bitcoin.
ETA: Sometimes I think Amazon's trying the Walmart strategy of inhuman growth and pushing through that strain & pain just to push out other B&M businesses in an area, then raising prices once a monopoly's established - which makes no fucking sense since the barrier to entry in a virtual supermarket is pennies on the dollar of producing a B&M supermarket. You can't keep competition out when barrier to entry is roughly equivalent to opening a freakin' lemonade stand.