The obvious story this week is Amazon Web Services's seemingly impressive financial results. From the WSJ:
The division’s revenue rose 49% to $1.57 billion in the quarter, which would put it on pace to generate more than $6 billion in sales this year, although Chief Executive Jeff Bezos in Thursday’s earnings report called it a “$5 billion business.” Analyst estimates for AWS revenue have ranged between $6 billion and $9 billion.
Amazon said operating income for AWS was $265 million in the first quarter, just $20 million above the year-earlier total. That follows the company’s pattern of spending nearly all the money it takes in on its various businesses. The operating margin was about 17%.
A lot has already been written about those headline numbers (with this piece perhaps being the best, deep analysis) and last year we heard AWS had over 2m servers and 1m customers. So instead I'll make a few points I think are not being discussed as much:
- Amazon should be congratulated for actually starting to report these numbers. No other cloud provider breaks down their financials in the same level of detail. It's true that Microsoft claims more revenue in "cloud" but defines the word very differently. There's absolutely nothing from Google, and IBM are worse than Microsoft with their announcements around cloud numbers. $3.8bn in cloud revenues in the latest quarter?
- Amazon continue to spend all their revenue and reinvest in the platform. They have been able to make the most of a huge first mover advantage but with intense competition from Google and Microsoft, there is no way they can reduce that investment. The problem is that Google and Microsoft both have significantly more cash available to invest in infrastructure. This will become even more important as margins for the IaaS products are squeezed.
- IaaS is a commodity. Compute, storage and networking are the core components that every cloud provider needs. Give how far along we are in the evolution of cloud providers, having a product portfolio that matches Amazon's core IaaS like EC2 and S3 is a massive barrier to entry and it's only really Google and Microsoft that compete here now. Other cloud providers like Rackspace and Softlayer have to find somewhere else to differentiate. This is a huge advantage for AWS because their IaaS offerings are very mature both in terms of features and ecosystem, which makes it very difficult for anyone to compete with.
- SaaS is not a commodity, and is a high margin business. It's a business well known to Microsoft and Google, and AWS have not stood still here launching all sorts of different SaaS products built on top of their IaaS - mail, calendar, source control, deploys, config management, file sharing, desktop, etc.
- This is all part of a plan to provide all of your IT needs - if you're already using part of AWS then it's trivially easy to sign up to other AWS products. Decisions to use IaaS tend to be made quickly by developers, then management buys in because it's already in use. Decisions to use a new mail system or document sharing platform seem to be more like management decisions, so it requires more effort to convince and migrate. Both Microsoft and Google started this way - with the SaaS products, and have since moved into IaaS - so that puts them behind in adoption because developers are already using IaaS from AWS.
- Margins will continue to fall in IaaS - the popular view is that it's a race to zero (although not really, in the same way that other commodities are low priced but not priced at zero!). This has to be offset by the high margins of SaaS, which is why all the cloud providers have such a range of products. It's also why the aggregate margin reported by AWS is a little misleading, because it combines both IaaS and SaaS.
Amazon have done very well over the last 10 years, and are well placed to continue innovating. They are far ahead of Google and Microsoft, but competition is only going to get more difficult. Google has really only been offering a compelling IaaS service (Compute Engine) for a year and their progress in that time has been impressive. Both Google and Microsoft have the core engineering capabilities to attack Amazon. I look forward to the day when they will split out their "cloud" revenue too!