"Cloud Wars" aren't over yet
Couple days ago The Economist published an article titled "Linux and AWS: Cloud chronicles". The article stressed the undeniable importance of Linux and Amazon Web Services (AWS) in the upsurge of cloud computing and went on to explain how network effects have helped both grow in popularity. Despite similar paths, article predicted that their future will diverge and Linux will "happily plod along" whereas "AWS could end up dominating the IT industry just as IBM’s System/360, a family of mainframe computers, did until the 1980s".
Even though AWS is clear market leader at the moment, the picture that The Economist painted doesn't fit market I see. If the article had been written a few years ago, the prediction would have been pretty solid but AWS is not as dominant as they were. To clarify where our disagreement lies, let me explain how their "model" works and why it is only half of the story.
Logic of the article
The (informal) analysis rationalizes the above mentioned prediction with the economies of scale and the network effects. Logic behind the economies of scale is very straightforward: Large operations bring efficiency and drive down the fixed costs. Network effects aren't as clear as in social network cases but they certainly exist. Large customer base generates lots of revenue that can be invested in R&D and large operations increase the know-how inside the company. Large user base also means that there are more potential recruits around and the internet is full of tutorials, cookbooks and answers to the very same questions you'll probably have. These channels do not have an immediate effect and hence it could be more appropriate to call them indirect network effects.
Semantics aside, in a market for homogeneous products the economies of scale and network effects (even indirect) will end in a (natural) monopoly. This is pretty much what the market used to look like and the reason why I said that the articles prediction would have been pretty solid few years ago. But by contrasting "S3 and EC2" to what AWS offers today, it becomes very clear that the demand today is too complex to warrant an assumption of a single homogeneous good.
Product differentiation and changing customer base: The other half
Sheer volume of different services and use cases means that it is very hard for a single cloud provider to be the best for everyone. Manufacturing company building an IoT platform has different needs than an online store and it will weight the importance of available components differently. Just by taking a look at the competition we can see that there are significant differences in market approaches. While AWS has been focusing on providing IaaS, Azure has focused on PaaS and more lately on analytical capabilities (I highly recommend this excellent post by Oskari for a more throughout discussion).
And it is not just the demand from the existing customers that is changing. Early adopters of the public cloud were largely media and tech companies whereas manufacturing, insurance and banking have been slower. This means that the size of the different market segments is changing over time. Extrapolating a trend from a self-selected sample rarely yields great predictions.
Summing it up
There are two opposing effects here. Economies of scale and network effects drive the number of cloud providers down. These effects alone would speak for an AWS monopoly. On the other hand, more sophisticated needs and diversifying customer base drives the number of service providers up.
At the moment Azure holds the second largest market share and a clear market position with PaaS and analytics offerings. Google however, is in an awkward position. It would have natural synergies in the tech and media sectors but they have already moved to AWS. Everyone knows that Google has the means to deliver but late entry to the market and lack of clear differentiation will probably cost them dearly.
Behind the big ones, everyone else seems to be struggling. HP threw in the towel and IBM is betting on hybrid cloud. At the end of the day, it seems highly unlikely that AWS will get so dominant that antitrust officials need to step in. My prediction is that AWS will remain on top closely followed by Azure. Google will burn enough cash to get 10-15 % of the market and Alibaba gets the wildcard. What do you think?