FORTUNE — When Amazon Web Services emerged in 2006, it was questioned far and wide. What, exactly, was the online mega-retailer doing dabbling in web infrastructure?
The move to provide services that have widely become known as cloud computing even puzzled some inside the company. “It was a big soul-searching decision,” recalls Adam Selipsky, Amazon (AMZN) Web Services Vice President of Product Management and Developer Relations. But the business, based on the ideas in a 2004 research paper, made sense to many, including CEO Jeff Bezos. As Selipsky tells it, most developers were spending 70% of their time working on tasks like setting up servers and managing databases, and just 30% of their time on their actual core business, taking orders say, or connecting users.
MORE: Amazon’s risky bet is paying off
Any doubts about Amazon Web Services (AWS) have long been buried underground. Just as other company gambles have succeeded — expanding beyond books, opening the marketplace up to third-party sales, the Kindle — AWS has paid off. Several years in, the seemingly left-field gamble that companies would rent data storage and computational power now just seems, well, brilliant. “It’s an interesting case study in how a company can take an objective and goal, externalize and not only drive a new business opportunity but refine an internal capability as well,” explains Forrester (FORR) analyst Brian Walker.
In 2010, AWS generated $500 million in revenues according to UBS Investment Research, a number that would balloon to $1.2 billion a year later. Now, by some accounts, AWS is a $1.5 billion annual business. AWS powers hundreds of thousands of companies, including Netflix (NFLX), Instagram (FB), and even NASA. They bank on Amazon for services like online database access, content delivery, and payments. Pinterest, the popular social network that draws over 17 million monthly visitors, uses AWS to manage over 400 terabytes of company data at any given time. Indeed, AWS has become so successful, that one analyst pegs its market share in the infrastructure as a service (IaaS) space at over 70%. (It’s nearest competitor, the San Antonio-based Rackspace (RAX), lags far behind, with just over 10% of the market.)
So what does the future hold for AWS?
MORE: Shedding light on the new Nook
First, expansion. Last year, AWS doubled the number of regions it operated in from four to eight, and Selipsky expects further geographic conquests over the next few years. While that means opportunities to build up its international client base, Amazon hopes it will be a boon for clients at home. If a company in the U.S., for instance, is trying to enter the Japanese market, Selipsky argues that transferring an application running in Virginia to AWS in Tokyo is simple. That would leave said firm focused on acquiring customers, not building infrastructure in a foreign country. Amazon is banking on this virtuous cycle boosting AWS’ fortunes abroad.
Walker however, suggests bigger changes for AWS may be in store. He argues that a spin-off of the massive cloud infrastructure business warrants consideration. Such a move may drive better value for fickle shareholders. More importantly, he believes it could help further differentiate AWS in the years to come, when the IaaS market eventually, inevitably heats up. Says Walker, “Certainly Amazon may be its largest customer, but people [may] want to ask: Does AWS belong as a standalone business versus one hitched to the back of an online retailing business?” Doing so might could alleviate concerns companies may have in the future about running their businesses on an infrastructure owned by a competitor.
MORE: Amazon’s knock-off problem (35 Shades of Grey, anyone?)
As much as analysts might war game the scenario or industry insiders whisper about it, Amazon says it has no plans for a spin-off. After all, Amazon runs on AWS, making a split complicated. What’s more, even as AWS has grown and supported an impressive ecosystem of other businesses, its revenues still pale in comparison to the monster $48.2 billion total Amazon pulled in last year. “There is a common global scale operations and infrastructure that is shared by Amazon’s e-commerce business as well as AWS,” argues Gartner analyst Ray Valdes, who says that behind-the-scenes synergy may be lost if AWS were spun out.
What about competing companies one day getting nervous about using Amazon’s digital pipes? Says Walker, “People will more likely think, if it’s good enough for Amazon’s online store, then it’s good enough for my small, little e-commerce business.”
