Rackspace (RAX) is a cloud hosting company whose stated mission is "To be recognized as one of the world's greatest service companies." In this article I express my doubts about the scalability of its business and whether its "cloud high" valuation is realistic. I recommend a short based on the impact that projected competition will have on its already thin margins.
First, a quick look at its financials for the four quarters leading up to 09/30/2012 (from Google Finance).
- Revenue: $1.24B
- Gross Profit: $884.7M
- Operating Income: $162.9M
- Net Income: $100.6M
- Market Capitalization at time of writing: $10.25B
Rackspace has more than doubled revenues since 2008 while keeping margins stable, which is probably why Mr. Market has rewarded it with a rich P/E of 106. More to the point, cloud computing has been "the next big thing" for years now, with global revenues expected to grow $240B by 2020.
It is unclear whether Rackspace is a hosting company or a consulting/service company. To see why, let's take a look at its business model (data is taken from Rackspace.com or its latest quarterly filing).
Rackspace's revenue comes from selling hosting services. It offers several products. Perhaps the most familiar one is its "Cloud Server," which is intended to compete directly with EC2 instances; it is priced almost identically to EC2 instances with the smallest configurations costing $.022 and $.06 / hour, and going up to $1.20 / hour.
Core to Rackspace's marketing image is that its hosting services are backed by the "Fanatical Support" of Rackspace employees, also known as "Rackers." It makes a point of saying that the number one differentiation between them and EC2 is its support.
As its customer base has increased, it has also increased the number of servers and the number of Rackers to scale. It has maintained the number of servers per customer at around 1 server / 2 customers, and the number of employees per customer at a steady 43 customers / employee.
Let's estimate how much capital Rackspace allocates to computers versus humans. The mean salary is $67,914 / yr according to salarylist.com, and $95,000 / yr according to indeed.com. At 4,596 employees, this works out to between $312M and $437M a year, not including benefits. The cost of a 1U server nowadays is around $1,000, which has to be rolled over about every 3 years. At 90,000 servers, this works out to around $30M a year in depreciation. Assuming each server uses about 500W / hour and $.127 / KWh, it probably spends about $50M a year on electricity, so $80M all told.
Another way to look at this is in terms of cost per customer. At 1/43 Rackers per customer, it pays between $1580 and $2210 / year / customer in terms of human power; at 1 server per 2 customers, it pays only $80 / year / customer in terms of compute power!
In comparison, Amazon (AMZN), which generated at most $1.5B in revenue from AWS in 2011 (i.e. comparable with Rackspace today), is thought to have had five times as many servers in early 2012, just for EC2. On the flip side, in early 2011 it was estimated that AWS was employing just 2,000 people. With similar assumptions as applied to Rackspace, this implies that it allocates $400M / yr for computing and between $130M and $190M for humans.
Will Rackspace's service-based business model work out? So far, it has been working out. It has increased revenues at around 30% a year since it went public, and its margins have remained more or less stable. However, I can't justify paying over 100 times earnings for its stock, no matter how big cloud computing is expected to be in 2020. Furthermore, its revenue growth has been slowing -- 27% this year as opposed to 32.5% a year ago.
All told, while I like the company's business and think it's a competent service provider, I have significant doubts about the scalability of its business and whether it merits a 106 multiple.
Unlike Amazon or Google (GOOG), which can simply increase the number of servers to scale with load increases, Rackspace has to screen, hire and train new employees. Put simply, if Rackspace grows 7 times (so that it will have a P/E of 15 at current valuations) and it continues its strategy of hiring employees to scale, then it will have 28,000 employees to manage, all of whom have to be just as trained and skilled as its current set of employees and be able to provide the same level of service that Rackspace currently does. It is not impossible by any means but a significant obstacle that it has to overcome in order to deserve its valuation.
And unlike consulting companies Redhat (RHT), IBM (IBM) and Oracle (ORCL), which offer consulting services across a wide variety of domains so as to cater to a large set of problems that its customers face, Rackspace only offers services for one very specific domain: setting up a custom cloud. This shows up in Rackspace's margins being much lower than those consulting companies, which is 7% vs. the 13%, 16% and 26% for Redhat, IBM and Oracle, respectively. Furthermore, since its margins are low enough already, advances in technology and price cuts by competitors will likely force it to spend more on computing per customer and drive its bottom line to zero or lower.
It is hard to short a stock which has been trending upwards as strongly as Rackspace. However, Rackspace faces significant competition from both fields that it is trying to play in -- hosting and consulting -- for example from Amazon's AWS on the hosting side, or Oracle's vSphere on the consulting side. If its margins shrink or revenue growth slows more, that would be a sign to me to short its stock or take profits on a long position.