> I listened to a talk by Amazon's CTO, Werner Vogels, where he was explaining that Amazon's entire strategy is to build an infrastructure for itself, then rent that infrastructure out to its competitors.

This was the game played by John D. Rockefeller. He had so much infrastructure built up by the time they split up Standard Oil, that each new state Standard Oil immediately had everything it needed to keep operations going and the breakup was essentially a bunch of paperwork. As a stockholder in all of the new firms, Rockefeller's net worth tripled, from ~300m to ~900 million.

To make this strategy work, you need mountains of cash. In his salad days, Rockefeller cultivated excellent relationships with bankers, who would stop him on the streets of Cleveland and ask him if he needed money. Eventually he decided to buy up Cleveland's entire refining industry, an event that became known as the Cleveland Massacre.

John offered three choices to rival refiners, cash, Standard Oil stock, or to get crushed. He leveraged his extensive railroad connections and infrastructure to bury competitors, for one thing he owned virtually all the oil tank rail cars, without which you couldn't sell your products because you couldn't move them efficiently.

The right choice was to have taken stock, which would have made you fabulously wealthy. If you were any good at the oil business at all, he would also offer you a job, which again would have eventually made you very very rich. He had an insatiable hunger for quality staff. It was never a good idea to bet against Rockefeller.

I listened to a talk by Amazon's CTO, Werner Vogels, where he was explaining that Amazon's entire strategy is to build an infrastructure for itself, then rent that infrastructure out to its competitors. That way, increased competition can actually fuel your growth. The talk was about APIs, but Amazon views APIs as a general term: "Fulfilled by Amazon" is an example of a non-technical API.

Likewise, I think Amazon's self-funded retail operations are largely a break-even enterprise. Other businesses have been successful with this -- Costco comes to mind -- but the scope of capabilities Amazon is looking to develop is staggering.

If they can build a completely vertically integrated retail platform from procurement to payment to delivery, basically they're operating a retail enterprise without taking any inventory risk on themselves. Which is what this article means by "capture a significant portion of US retail": if they can provide merchant services to investors, it turns Amazon into basically an investment bank. Amazon takes a retailers money, turns it into inventory, sells it, takes a cut, then returns money. The only inputs and outputs into the Amazon machine are money, and it turns retail operations into an investment product.