A November to Remember at AWS
November was an action-filled month at Amazon (NASDAQ:AMZN) AWS. The company held its re:Invent user group meeting in Las Vegas and used that forum to announce a whole spate of new service announcements. The company announced a series of price reductions which took effect at the start of the month - sort of a premature Christmas for most of its users. And perhaps of most importance to some observers and stakeholders, AWS announced its ambitions to "move up the stack" and there were, what some might feel to be, indications that Salesforce (NYSE:CRM) could be the favored partner in that dance.
For some companies that might be enough activity for a year. At AWS, it all took place within a month, which I propose to review. I'm not going to review Amazon Go. Others have done so and have been impressed with the concept.
To be concise and upfront, I recommend that readers add to or establish positions in Amazon. I'm happy to have the debate about the value and future of the company's e-commerce operation, but this article is going to stick to looking at recent happenings at AWS. I have recommended Amazon shares since I first wrote about them on this site back at the start of April. The shares are up 27% since back then compared to a 9% gain for the tech software index. The shares have dropped about 10% since the high they made before Q3 earnings were released. That drop, I believe, offers investors a decent entry point.
This morning Citi removed Amazon as one of its top internet picks and replaced it with Google (NASDAQ:GOOG) (NASDAQ:GOOGL). I haven't yet seen the report so I can't comment on its insights or lack thereof in evaluating if there is something in the firmament with which I should be concerned. The shares have shown little reaction thus far.
When it comes to evaluating AWS against Google Cloud, I still like what AWS does better from an investment perspective. Google has an aggressive price/performance. It also has some strong service offerings. But it can't match the geographic reach of AWS and it simply has little chance of developing all of the services that AWS offers and will continue to offer.
I'm not going to specifically try to debate those who feel that AMZN shares are overvalued. It is, I think, a somewhat fruitless debate. No one knows how to quantify the worth of unicorns - this is a company that is coming up with unicorns consistently, be the domination of the Indian e-commerce market or upending the way people shop for groceries. Can anyone really create a DPV model for the profits to be had from dominating the e-commerce market in the world's single largest country, India, by population? I can't and don't think it is fruitful to guess. (Frankly, Amazon upended the way I shop for groceries a few years ago when it introduced Prime Pantry and I discovered I could avoid having anything more to do with the purchase of detergents and cleaning supplies and bottled juices other than making a few clicks).
For those so inclined, this site carries negative articles regarding AMZN and its prospects all of the time. It tends to focus on traditional valuation metrics and financial analysis and often scream about the company's hyper valuation. It should be obvious by this point in time that Amazon shares are simply not for everyone. If one's perspective is short term, if one wants dividends, if one desires Graham/Dodd valuation metrics - these shares are not for you. I think the trick here is to acknowledge that there are many different classes of investors and to understand how Amazon is optimized for the predilections of growth investors with a long time span.
But the net of the "November to remember" for AWS is another set of signposts on the path to significant returns to Amazon's shareholders. AWS is valued in different ways by different analysts. It has such a high growth rate at scale and is seeing profits rise so rapidly that it presents significant challenges in using standard valuation tools. What is the terminal growth rate to use? What is the CAGR prior to the terminal growth rate? How long will it be until AWS reaches a steady state growth? How profitable will AWS become? And finally, the biggest unknowable, what is the appropriate weighted average cost of capital to use in the equation? I know - absolutely - I have no way of knowing those things in advance. If others have figured out where those answers lie - come ahead. I think I can say, however, that the headlines from AWS last month have all contributed to my belief that AWS is undervalued and that it remains, even more than some of the other disruptive technologies of this company, a jewel still quite undervalued in the consensus thinking.
Once upon a time, and really not that long ago, school children in what was then known as the British Empire were taught to recognize a particular shade of pink. Within that zone, which covered 25% of the world's surface, the writ of the British Empire ran and the sun never set. By 1931, when the Statute of Westminster came into effect, the case was harder to make (Canada, Australia and New Zealand were really independent from that date) and of course, after WWII - well… mapmakers had to figure out what to do with surplus pink ink.
At the moment, the sun may set on AWS. It doesn't have any regions in either the Pacific or the Atlantic and the company does not have either a region or an availability zone in Africa and the Middle East. Lest I be accused of hyperbole, however, the fact is that with 14 operating regions and 38 availability zones, AWS is a behemoth dwarfing its competitors. In a recent article published in Fortune, Gartner says that the server capacity of AWS is "many times the size that of the other providers in the market." The company is on track to release 1,000 new service capabilities this year, up from 722 new services last year. I'm absolutely certain that I couldn't tell you what all of these services do - maybe Mr. Jassy, CEO of AWS, might be able to figure all of them out.
Smothering the competition, cosseting the users
It is really two sides of the same coin. From time to time, I read about how the many competitors in the cloud space are either taking or will take share from AWS. It is clearly true that AWS with its 55% growth rate is growing more slowly than either Google or Microsoft (NASDAQ:MSFT) which are enjoying triple-digit growth. It is also true that so far as the statistics can be relied on, AWS has more than twice the market share when compared to MSFT and 5X or more the market share that Google has. AWS users simply have more choices in terms of locations and functionality than they can find in choosing competitors. And that advantage is likely to continue to stretch.
I will just pick, somewhat at random, a couple of interesting service announcements from the conference. One service offering is called the Snowball Edge. Using the service/product, businesses can collect IoT data and move it to the Amazon cloud, which would analyze the information for actionable intelligence. Needless to say, the service is going to compete against the offerings of Cisco (NASDAQ:CSCO), Hewlett Packard Enterprise (NYSE:HPE) and Juniper (NYSE:JNPR) as well as the larger server vendors. The edge includes a new technology called Greengrass (not the sturgeon king in New York). It is called serverless computing and it ought to be obvious as to the impact the technology may have in the server world.
Other announcements considered significant by technical types who write about such things include new instance types, a further push into the hybrid cloud, which can be facilitated by Snowball Edge, and more machine learning options that are an attempt by the company to catch up in AI.
The point in talking about the services is not to catalog the offerings of AWS. The point is to try to explain to readers and investors one singular advantage AWS has in the field. It is as close as it is possible to be a one-stop shop for cloud computing and it is drawing away from competitors in terms of the functionality that it offers. The virtuous cycle, in which AWS is larger and, therefore, offers more services that attract more customers that allow it to become larger still, is alive well and working nicely.
As Rich Sutton, VP of engineering at Proofpoint (NASDAQ:PFPT), said, "I don't perceive there to be too many instances types, they all have their tradeoffs, but unlike the old hardware days where you had to lay out Capex to figure out which is the best for you, you can just try the stuff out." That is the kind of moat and advantage that many writers who like to down-value what this company has created, are simply unaware. No, not all cloud offerings are alike and Amazon is not only the largest but it offers its users more functionality.
Some of the things that AWS is offering are by way of being overkill. Snowmobile - living inside a 45-foot-long semi that was driven on to the floor of the show as a demo, allows users to transfer exabytes of data to AWS with a 10 Gbps line. That is probably more capacity than most users might need, but it is just another niche service that might weigh on the scale when users consider developing enterprise relationships. Quoting another consultant, Mike Kavis, VP and principal cloud architect at Cloud Technology Partners,
"Even though they (AWS) have more feature function than anyone, those areas where they do have gaps, they're starting to address. They're starting to remove the excuses for you to go to other places."
I am sure if I looked hard enough I would find other users with positive things to say about MSFT and Google cloud offerings. But I think that it is difficult to make the case that AWS does not have some significant competitive advantages that will not be easy to replicate in the foreseeable future.
Moving up the stack - A chimerical hope or something that might be another disruptive piece of the AWS strategy?
Obviously, AWS loses many deals and doesn't compete in others. There are many reasons why that happens but one of the more important of them, I think, is that until now, Amazon has been unable to offer a full stack solution. What that means in the real world is that a user migrating to the cloud can utilize Oracle (NYSE:ORCL), SAP (NYSE:SAP) IBM (NYSE:IBM) and perhaps, most importantly, Salesforce to get both an enterprise cloud and applications that are specifically tuned to run on that cloud. Amazon offers the cloud, a set of database and some security options but users need to do their own integration.
That is now starting to change and may wind up significantly disrupting the industry in favor of AWS yet again. About six months ago, Salesforce announced it would be building some of its newer products to run on AWS as well as on its own cloud. During the latest Amazon event, the companies announced that all of its current products would be running on the Amazon cloud in Canada a few months. In addition, Salesforce is buying $400 million of AWS infrastructure over the next four years to make that happen. It is a PR coup for AWS no doubt and for the first time, they will be able to offer a stack that includes competitive applications. The effort involved for Salesforce is non-trivial. Up till now, it has used Oracle (ORCL) databases running on what techies call "bare metal" servers. Amazon uses virtualized servers and it is not a simple undertaking for Salesforce to move its products to Amazon.
Next summer, at least in Canada, users will be able to buy the same solution running either on Amazon or the Salesforce cloud for the same price. The newer Salesforce products, particularly, its IoT offering, have always been designed to run on Amazon's cloud. In addition, it was announced that AWS and Salesforce would enable Herouku, a development tool that is used by Salesforce customers to run on Amazon's Virtual Private cloud.
Looked at holistically, over time, all of the Salesforce solutions are going to be available from Amazon, running on Amazon servers with the same pricing and the same functionality as will be offered by Salesforce.
Salesforce obviously thinks that there is a large market it is not properly addressing and is spending a very significant sum of money to enable its products to run on Amazon. What the parameters are of the opportunity are not known, but a company the size of CRM spending $400 million on porting its apps to Amazon is a significant statement.
I think it is fairly obvious that conflicts are almost inevitable when Amazon and Salesforce offer the same solution, running on the same hardware at the same price. Will that actually happen or is a harbinger of something else? Needless to say, speculation is likely to emerge regarding the something else - and the something else would be an acquisition of Salesforce by Amazon.
Needless to say, no one has troubled themselves to tell me. But for those with only a moderately long memory, it was not that long ago that Microsoft - and others - were said to be pursuing an acquisition of Salesforce at what at the time seemed to be significant premiums. 18 months or more have passed since those days. On an EV/S basis, Salesforce shares are much less expensive than they have been.
Microsoft Dynamics, particularly, Dynamics CRM, has become a real competitor against Salesforce in the cloud. Amazon and CRM have become partners at a significant level and Salesforce has made significant investments in order to fund its software on Amazon servers. And Amazon says it wants to "move up the stack." Apparently, Amazon's Redshift data warehouse is being integrated with Salesforce technology to promote analytic solutions. And a new set of tools will help developers incorporate voice recognition that Amazon developed for Alexa into CRM software.
I do not purport to run these businesses or to know how to do so on the ground. I don't have the opportunity to interview Keith Block or Marc Benioff or their counterparts at Amazon and if they told me something substantive, I couldn't use it in any event. But I think it is at least reasonable to think about what the partnership is going to mean to all of the companies involved going forward. Amazon is a significant gainer, just on the surface. If it can strike a deal to acquire Salesforce, it would be one of the more audacious strokes to counter Microsoft's grand strategy and would lead, I believe, to a host of synergistic benefits both apparent and not so apparent.
Pricing - the 600-pound gorilla in the room
Late last month, Amazon announced a series of price reductions on various components of its service offerings. It has reduced the price of both standard storage and Glacier storage. While determining the average price cut for AWS is not simple, the reductions in the US for the company's storage offerings range from about 21-24% and the cuts range to 28% outside the US. The cuts on Glacier cold storage are greater. Other service offerings such CloudWatch Metrics have also seen noticeable price reductions of varying percentages.
There have also been price reductions on various kinds of instances, although of a much smaller magnitude as well as a reduction in snapshot pricing. I think this series of reductions brings the total number of price cuts to 55 since AWS formally launched 10 years ago.
If AWS is growing at 55%/year, why does it need to cut prices by a noticeable amount every few months? Some people believe that Amazon is engaged in a price war. Some people believe that Amazon growth is declining - and of course, ultimately that will be true. At the start of the year, AWS growth was only supposed to be 45%. To me, a price war implies that the companies involved are selling their products/services less than the marginal cost of producing the products/services. That isn't what is happening here.
Amazon sells services in one of the more dynamic spaces in the world when it comes to technological innovation. The cost of the components that make up storage for Amazon, and for everyone else, decline dramatically year on year. The cost of everything else that makes up the large data centers which are the equivalent to Amazon's cost of revenues decline consistently as well. The Snowball Edge serverless data appliance, which has significantly lower TCO than traditional products, has obviously seen its initial incarnation in data centers that Amazon operates for itself.
Amazon is less than transparent in revealing most of the operating metrics for AWS. There isn't a separate P&L breakout or even a proxy for most operating metrics. What we have seen, however, is that operating margins reached a record 32% last quarter. How much of that was because of increasing gross margins and how much of that had to do with opex leverage is simply unknown. My guess is that there is some of both, but I would be inclined to believe that gross margins have been driving these increases. At some point, higher gross margins imply that the company is either ignoring price elasticity and that it is presenting a larger target for potential competitors. While not going into the subject here, price elasticity in cloud services is enormous as it supplants on-premise workloads. And hence, the price cuts in the wake of ever growing demand. It is the right strategy for AWS - it is the exactly right strategy.
The other day, a commentator on this site posted an article about the potential for Oracle to beat AWS. To a certain extent, the idea that Oracle is going to beat any large competitor in the business of selling cloud capacity is an audacious one. The author has pointed out that in an effort for Oracle to catch up it has chosen to acquire many companies with cloud-based solutions. Clearly the acquisition of NetSuite dwarfs everything else that Oracle has acquired in terms of size. Some of the other smaller acquisitions are aimed at improving Oracle's positioning in an offering that competes against Amazon.
I have my doubts about how the NetSuite merger will work for Oracle. Making mergers and finding synergies and improving competitive positioning are simply not all the same thing. While Oracle bought some technology, Amazon introduced 1,000 offerings of various kinds - not all of them terribly relevant to a broad customer set - but all part of a strategy of smothering competition.
A couple of months ago, during the course of its annual sales meeting, when CEO Ellison was proclaiming his war with AWS, he mentioned, and not in passing, that the new Oracle IaaS offerings would be far cheaper and more functional than the Amazon offerings. Mr. Ellison is notorious for those kinds of statements. Braggadocio surely, personality dysfunction, maybe. But the fact is that none of what has threatened has seen the light of day yet and may never do so in the form investors might imagine. If I were Oracle management, I would be far more concerned about the potential for an Amazon/CRM joint offering that offered the advantages from both vendors. I think that the Amazon/CRM combination offers far more competitive challenges for which Oracle has no obvious answer (although Mr. Ellison will surely try to have an answer when Oracle reports earnings next week) than the challenge that Oracle poses to Amazon. Oracle may or may not have produced pleasing numbers - those numbers are almost certainly not a function of Oracle's successful competition with Amazon and I doubt that Oracle will ever be able to replace Amazon because of any advantages it may develop or acquire.
Amazon has reduced prices and some investors may think that is a representation of either market share loss or slowing demand. Indeed, Google maintains that it is still lower-priced than Amazon. That may be so - there are services these days that are used to analyze the pricing of all of the service offerings users can buy. The pricing for different offerings has become so complex that it has become difficult for users to determine who really has the most attractive price/performance for a particular workload.
No one is going to suggest that Amazon can or would be able to use its cost advantages to drive its competitors from the field of clouds. But Amazon is doing what is prudent and optimal to maintain its market share and growth the market while maximizing its profits.
Other analysts have suggested that all things being equal, the current price reductions will reduce reported revenue growth from AWS for a quarter or two. All of the current customers get a price decrease. The visible impact of the price action will be small this quarter as it only affects December revenues. I would be surprised if AWS growth rate and operating margins didn't show declines in the first couple of quarters of 2017. For all of 2017, however, I expect that revenue growth for AWS will be greater in total dollars than the $5 billion of revenue growth likely this year. And I think that operating margins are going to end up higher for the full year as well.
I tried to present, at a high level, the success AWS has had and is likely to have in the future. It is one of the keys to the longer-term profitability of AMZN. The company has made a number of service announcements at its re:Invent user meeting that extend its lead in terms of the technology it offers. It is forging a significant partnership with CRM that is leading to combined offerings that fill the stack for the company. And it remains a price leader, willing to trade short-term margins for growth and to price in order to maximize market share along with profitability.
It is very hard to use analytical tools to value this unicorn. I think it is surely worth more than the $150 billion valuation that has been written about by commentators. I think it will grow more rapidly and for longer than is assumed by most DPV models and it will ultimately achieve higher margins as well and hence generate more cash. Investors ignore the news from Amazon's "November to Remember" at their peril in terms of finding the best of the large-cap growth tech names.
Disclosure: I am/we are long AMZN.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.