Summary:

• In FY16, AWS' operating margins dwarf the e-commerce business' -- 25% vs. 3% -- contributing 74% of AMZN's operating income compared to just 9% of total revenue.
• The increasing number of higher-layer, PaaS offerings and their growing enterprise adoption provides another strong growth vector for AWS.
• The probability and magnitude of the things that could go right for Amazon are larger than that of those that could go wrong.

Our inclusion of Amazon as a Samadhi 4 holding (our four highest-conviction holdings, with weights ranging from 5% to 10%) can be justified almost completely on the basis of AWS (Amazon Web Services), the company’s suite of market-dominant cloud offerings in the IaaS (Infrastructure-as-a-Service) and, increasingly, PaaS (Platform-as-a-Service) segments. That is not to say there isn’t value in Amazon’s other segments (many of which increasingly find synergy with AWS), but that they factor in substantially less to our overall thesis.

The Shift to the Cloud Drives AWS

In FY16, AWS' operating margins dwarf the e-commerce business' -- 25% vs. 3% -- which allowed the segment to contribute 74% of AMZN's operating income compared to just 9% of total revenue. At 55% y/y growth, AWS also grew top-line nearly twice as fast as e-commerce. We believe that, as a key beneficiary of the enterprise migration to the cloud, AWS will grow by 58.5% annualized over the next five years and continue to be one of the largest cash cows in the tech sector.

The confidence we have in our 58.5% revenue CAGR over the next five years is bolstered by the fact that we are still early in this enterprise cloud migration. Estimates from leading industry and equity research firms indicate that only 18% to 22% of all application workloads (which include services across the three layers of the cloud stack: IaaS, PaaS, and SaaS) had moved to the cloud in 2016. These same estimates indicate that, by 2022, more than 50% of application workloads will be running in an external cloud service provider’s datacenter.

Enterprises are also likely to move increasingly mission-critical – and therefore less price-sensitive -- applications to the cloud, which would offset impending price-based competition from newcomers like Google and Oracle (leading to increased value from AWS’ tiered service levels).

As an aside, though Oracle’s database software serves as the basis for many of the most mission-critical applications that enterprises run, Oracle itself does not have nearly as much experience in operating these mission critical applications as its customers do.

Amazon, on the other hand, has over a decade of operational IaaS experience to bring to the table. In fact, Amazon already provides the computing infrastructure for a growing number of SaaS vendors and other types of Internet-based businesses – i.e., those that rely on IaaS the most. For these customers, most of which are rapidly growing their own businesses and therefore their consumption of AWS services, the software they run on AWS isn’t just mission critical because it helps them manage their business, it literally is their business. Some examples of notable SaaS/Internet-based businesses running on AWS include: Netflix, AirBnB, Pinterest, Zillow, Twilio, Workday, and even Salesforce.com for certain applications, which recently purchased $400m worth of AWS capacity for use over the next four years.

We also believe the increasing number of higher-layer, PaaS offerings and their growing enterprise adoption provides another strong growth vector for AWS, which offers a broad and growing PaaS menu that includes multiple types of database-as-a-service offerings, analytics offerings, wide variety of developer tools, premium security features, etc. AWS’ PaaS suite seems to add some new service almost weekly. These services also do not require AWS to exclusively dedicate as much data center capacity as the lower-layer IaaS offerings and are therefore significantly higher-margin, which should fuel even greater operating margin expansion as they become a higher percentage of the overall mix.

Our Scenario and Expected Return Analysis.

We utilize a probability-weighted five-scenario model, with a Weighted Average Expected Return (WAvER, as we call it), to shape our projections and expected returns. Our WAvER projections for Amazon yields an expected return of ~33% annually over the next five years, which is a little higher than our base case, and assumes the following:

1. AWS performs as we expect with a 5-year revenue CAGR of 58.5% and op margins moving from 25% to 39% during that period;
2. Modest performance from Amazon’s core e-commerce business with a 5-year revenue CAGR of 16.1% and op margins moving from 1% to 8.5% during that period;
3. No material revenue from its various hardware products or from any other speculative R&D projects;
4. A terminal multiple of 28 times 2022 earnings.

After working through the plausible downside and upside scenarios, we believe the probability and magnitude of the things that could go right for Amazon are larger than that of those that could go wrong. Even in our “Crazy Down” scenario, to which we give a generous 21% probability of occurring, the stock produces a 3% annualized return over the next five years.

Noteworthy assumptions in this Crazy Down scenario include a highly conservative, steep revenue growth deceleration from 27% in 2016 to 14% in 2020, plus a convergence to what we've termed “reactive operating leverage” (our term for a phenomenon where management accepts the reality of slower top line growth, realizes that marginal spending on R&D and Sales & Marketing will produce a lower than marginal impact on growth, and therefore reacts by limiting or even reversing spending growth, thus generating higher EPS growth than revenue growth – at least until operating margin approaches the realistic limit for the business model of the company in question).

Given that Bezos, ever the long-term optimist, has historically eschewed reactive operating leverage in favor of funding new projects, we believe that this assumption is prudent. Having said that, there is still substantial “natural operating leverage” potential inherent in Amazon’s business model—especially in AWS. Even when accounting for the wide and wacky cone of outcomes inherent to the tech sector, we have strong conviction that, over the long term, there are few large cap stocks with Amazon’s attractive risk-reward profile.