Should Amazon Spin Out Its Cloud Operations?
Eventually it will be inevitable, says Tim Horan, an analyst with Oppenheimer, who shared the idea in a research note with clients today, because the business will interfere with other higher-priority businesses as it grows. He reckons that AWS — commonly known as Amazon’s cloud services business — is bringing in somewhere in the ballpark of $2 billion a year in revenue, which would amount to about 3 percent of sales, and he projects that it will grow to $10 billion within about three years. As yet, Amazon has not disclosed the real figure.
The problem, as Horan see it, is that when retailers turn to cloud computing, they’ll have little choice but to call Amazon, which they see as a competitor. There are other options, ranging from Hewlett-Packard’s burgeoning cloud services unit to another one at IBM, to Rackspace, Joyent, and Verizon’s Terremark, to name a few. Using Rackspace for comparison, he estimates that Amazon’s AWS could be worth as much as $101 billion by 2018.
The spinout notion seems to ignore one fundamental truth about Amazon on AWS. Amazon itself runs its retail and media operations on AWS, so the revenue it generates from selling computing-for-hire services of every stripe subsidizes and indeed minimizes the computing overhead costs associated with everything else that Amazon does, including retail and media streaming. That’s where the logic of a spinout, at least to me, breaks down. If it were to happen, wouldn’t the cost benefits that Amazon gets from running on the internal AWS infrastructure be lost?