1. I'm not so sure about that. According to the most recent financial statement[1], AMZN hasn't generated much in the way of net cash from operations over the past 12 months. Worse yet, they're spending an amount in excess of that on investments (some of which is really capitalized operating expense, like internal software development, that companies like MSFT expense directly). In fact, for the 12 months ended 6/30/2013, the only reason ending cash was greater than beginning cash was because they undertook financing transactions (selling stock, issuing debt, etc.) See the section that discusses "free cash flow," which is in itself a non-GAAP measure, as it omits non-avoidable expenses like operating lease payments, principal repayments on debt, etc. In other words, even that measure doesn't really represent distributable cash.
2. Unlike SaaS companies that sign multi-year contracts as their main source of revenue, the vast majority of Amazon's revenue is products, for which revenue is recognized almost immediately. Over time, then, the cash the company collects should roughly correspond to its GAAP income, unless services that are paid up-front for consumption over time increase as a % of its revenue (e.g., reserved instances, annual subscription fees, etc). In other words, it is unlikely to profitlessly generate cash; the two should rise or fall together.
3. Amazon is a fearsome competitor. But ORCL makes vastly more money. In fact, when Walmart's revenues were the size Amazon's are today (in the early 80s, roughly, adjusting for inflation), it was far more profitable then than Amazon is today--despite having to operate stores and truck fleets.
[1] http://edgar.sec.gov/Archives/edgar/data/1018724/00011931251...