You can probably rattle off the top enterprise software vendors without thinking: Microsoft, IBM, Oracle, and SAP. According to the best estimates I can find, those four companies together racked up close to $140 billion in software revenue in 2015, led of course by Microsoft and its well-known offerings.
Our cloud future will feature a different foursome -- Amazon, Microsoft, Google, and IBM -- and although public cloud revenues remain a small fraction of those driven by software, growth is by leaps and bounds across the board. AWS, whose long lead seems to grow and grow, pulled in more than $7 billion in 2015, a year-over-year expansion rate of around 80 percent.
Microsoft’s cloud business appears to be jumping as well, with an analyst at the firm FBR Capital Markets predicting that Redmond will break $8 billion in cloud revenue for 2016, up from an estimated $5 billion this year. That number includes Office 365, however (which may be cloud connected and cloud delivered, but you can’t really call it SaaS).
Based on purported revenue alone, IBM is next in line, since its last quarterly earnings report claimed $4.5 billion in annual public cloud revenue, a 45 percent jump year over year. But like Microsoft -- which was recently chastised for fudging its cloud numbers by none other than ex-CEO Steve Ballmer -- IBM has a history of inflating cloud revenue for the enjoyment of analysts.
Google doesn’t even break out its cloud revenue, but one source, The Information, estimates it to be a mere $400 million for all of 2015. Sounds about right, because until recently, it’s been hard to determine whether Google had an enterprise cloud strategy at all.
Then, two months ago, Google Technical Fellow Urs Hölzle predicted that Google’s cloud business could outpace its advertising business in five years. To put that in context, Google made around $65 billion in advertising in 2015.
How is Hölzle’s conjecture even remotely possible? Because in the public cloud business, infrastructure is everything. Check this recent post by InfoWorld contributor David Mytton: “Global location wars: Amazon vs. Microsoft vs. Google.” Google is No. 3 in its global coverage, but the point is the company already boasts an enormous cloud data center footprint thanks to its search/advertising business.
Note that David’s map charts public cloud availability regions, not capacity, but there’s a reason why he left out IBM: Most of IBM’s global presence has come from a buildout of SoftLayer, which lacks the platform services offered by the other three players. The IBM Bluemix PaaS has a rich bundle of services, but as far as we could determine, Bluemix as a public cloud offering is currently available only in the U.S. South, the United Kingdom, and Australia.
Given IBM's traditional approach to business, this makes a certain degree of sense. IBM professional services can build out whatever its customers want on its SoftLayer infrastructure, with Bluemix availability gradually ramping up over time. Meanwhile, many Bluemix deployments will be on-premises, as IBM plays a hybrid long game. The other three players have more explicitly dedicated themselves to delivering public cloud self-service.
Amazon established that model, which accounts for its huge lead -- although enterprise customers remain a small slice of its clientele. Microsoft has the unique advantage of a huge presence in the enterprise data center with Windows Server and System Center, which (with the help of Azure Pack and Azure Stack) it’s already using to foster a hybrid architecture for customers, with the goal of making Azure cloud a natural extension of on-premises customer infrastructure. That smooth on-ramp is part of the reason behind the bullish predictions that Azure may soon overtake AWS in the enterprise cloud competition.
Google has the most room to grow. It has a head start in the race to support production container deployments at scale, thanks to pioneering work on the Linux container spec and experience spinning up billions of containers per week -- along with the recent launch of the Cloud Native Computing Foundation, which may well develop an effective hybrid play. In a December 2015 review, InfoWorld’s Peter Wayner found the Cloud Platform as it stands to be a flexible, elegant offering.
Will Google stay committed and figure out how to accommodate enterprise customers? Will Amazon offer enterprises an easier way to go hybrid with AWS? Can Microsoft sync the ongoing development of the Azure public cloud and the Azure Stack effectively? At the least, we’ll see a glimmering of answers to these and other questions in 2016.