In my previous post, I explained the three different types of Cloud computing, SAAS, PAAS, and IAAS. I also argued that lumping all flavors of Cloud computing together could be misleading to investors. For nontechnical people, I suggest viewing this short video clip here to get an overview of Cloud computing for IT operations.
In this post, I will specifically discuss Infrastructure as a service (IAAS) and describe Amazon’s (AMZN, Financial) multiyear lead in the industry. Amazon’s IAAS platform is named Amazon Web Services (AWS). AWS first delivered Amazon S3, a storage service, in the spring of 2006 and Amazon EC2 compute in the fall of that year. In its Q2 earnings report this year, Amazon disclosed that AWS generated $1.82 billion in revenue and operating income of $391 million and is projected to generate $6 billion revenue for the entire year.
You can see how AWS ranks in comparison with its competition via Gartner’s IAAS Magic Quadrant released in May.
Quantifying AWS's lead
Per the Gartner report, “Magic Quadrant for Cloud Infrastructure as a Service” for Microsoft (MSFT, Financial), which is second place in IAAS market share, “[Its] Azure [offering] is growing rapidly ... with more than TWICE as much Cloud IaaS compute capacity in use as the aggregate total of the remaining providers in this Magic Quadrant (excluding market share leader AWS).
Furthermore, “[AWS] is the overwhelming market share leader, with over 10 TIMES more cloud IaaS compute capacity in use than the aggregate total of the other 14 providers in this Magic Quadrant." In other words, the competition is really, really far behind Amazon especially vendors not named, Microsoft.
Understanding AWS’s scale
There’s a really good article by Enterprisetech.com here. Notable highlights:
- Each day, AWS installs enough server infrastructure to host its etailing business from back in 2004.
- AWS has added enough capacity to support a $2.55 trillion online retailing operation. For reference, U.S. 2013 GDP was ~$16.8 trillion and China’s 2013 GDP was ~$9.2 trillion.
- From Gartner, “AWS groups its data centers into 'regions,' each of which contains at least two availability zones. It has regions on the East and West Coasts of the U.S., and in Germany, Ireland, Japan, Singapore, Australia, Brazil and China. It also has one region dedicated to the U.S. federal government.”
- Customers architect their applications to run across different availability zones.
- Amazon scales up the capacity in each availability zone by adding whole new datacenters.
- A typical Amazon data center has at least 50,000 servers and sometimes more than 80,000 servers.
- Timothy Morgan from Enterprisetech.com estimates that AWS could have upwards of 5.6 million servers worldwide.
What are the benefits to customers?
Data centers are work intensive and costly to build. Items may include but are not limited to real estate to house the hardware, servers, software, networking equipment, power equipment, cooling equipment, etc. On top of that, you’ll have to hire people to manage the facilities and keep the data center temperature controlled, consultants to implement, support staff to maintain the equipment, etc. To get an idea of itemized costs you can view a data center calculator here. Amazon knows managing data centers is not most companies’ core competency, and it imagines a world where customers plug into its infrastructure and only pay for the capacity it uses.
Examples of why customers may find AWS or any other IAAS provider attractive:
- A retailer will likely have much more web traffic during the Christmas shopping season than in March, and it would be inefficient to have extra computing equipment laying around just for one time of the year.
- Companies that need a global presence may not have the funds to build data centers around the world.
- Startup companies competing to be first to market would need raw computing power immediately to develop and test their applications.
- You can see short video clips of how organizations like Major League Baseball are using AWS for motion tracking or how Siemens (SIE, Financial) is using AWS to improve cancer treatment here.
One size does not fit all
Despite being able to boast recognizable clients such as Samsung, Netflix (NFLX, Financial), Pfizer (PFE, Financial) just to name a few, AWS will not be an ideal choice for everyone. In 2013, a Seattle-based marketing company named Moz dumped AWS complaining about cost, product stability and service. Gartner points out that, although AWS has the most robust offering it might not fit for customers with very unique requirements. AWS’s support staff may also get stretched as it must prioritize its customers and smaller accounts may be affected
Final thoughts
At $1.82 billion in revenue for Q2’15, AWS is just a fraction of Amazon’s overall revenue of $23 billion for the quarter. Year over year, AWS revenue grew 81% and operating grew 314%. From this Forbes article we get the following graph where IaaS is forecast to be ~$45 billion by 2018
Amazon’s FY14 revenue was $89 billion, a 20% increase from 2013. If I’m extremely generous and assume its revenue grows 25% from 2014 to 2018, I would get forecast revenue of $217 billion for 2018. If I give AMZN a 5% net income percent for 2018 which is a benchmark that it hasn’t achieved in 10 years, then I get net income of $10.9 billion. They would need to drastically cut R&D and SG&A expense since last fiscal year, they actually lost money. At today’s price of $527.39/share and the current number of shares outstanding, I get a FY2018 forward PE of 23. For perspective, Walmart’s (WMT, Financial) revenue for last fiscal year was $485.7 billion, and its net income margin was 3.37% so my rosy FY2018 scenario is possible. In the end, even with AWS’s dominance, I consider Amazon’s shares to be overpriced. You can see my thoughts about how the cloud affects IBM here.
