Re-visting my AWS Thesis and Valuation $AMZN

Reading Time: 3 minutes

I published an AWS piece in March 2019 that is counter to what most people are saying. My note could be boiled down to 3 thoughts: AWS is a good business with great tailwinds, but:

  • It’s becoming highly competitive, with Google, Microsoft, Oracle, IBM and others all gunning for a piece of the pie.
  • Pricing, and therefore margins, will likely go down in the future . In fact, AWS proactively lowers pricing for customers. This is not a sign on a good business, but instead a sign of a commodity. You become the low cost producer by having the highest volume, so you need to incent players to stay and leverage your fixed costs.
  • I didn’t think it was being mis-priced in Amazon’s stock. In other words, it wasn’t a hidden gem. You’ll see from this post that I still think that is true.

I’m reposting my initial projections for the company below and we’ll discuss how they’re performing to those expectations. I came up with a value of $165BN.

Maybe I was being a little harsh on the out years, especially for margins. But guess what? AWS was somewhat in-line with this estimate in 2019, with $35BN in sales, $9.2BN in Op Income and $17.4BN in EBITDA.

2020 is supposedly the golden era for AWS with COVID-19. Tons of start-ups are growing rapidly with everyone shifting to the cloud and AWS is the backbone they are built on. As those companies grow, they scale up the resources needed from AWS so AWS makes more money.

YTD 6/30/2020, AWS did $21BN in sales and $6.4BN in EBIT.  If we annualize that, its $42bn in sales and $12.8BN in Op income for ~30.5% EBIT margin. So they are behind on sales actually, but the margin is holding in there. This is from higher utilization needed during the pandemic (which helps fixed cost leverage), but it looks like the company also changed its depreciation schedule, so I would need to see D&A to get a real sense of what true cash margins are (and they don’t disclose D&A by segment outside of the 10-k). They actually called this out as helping the EBIT margins.

For those that may think AWS can keep up its amazing growth, that’s not so. Its clearly decelerating, but it is on big numbers so that is expected. By contrast, Azure grew 50% in this same quarter (though is about half the size of AWS).

So maybe I’m wrong on the AWS valuation. I do think I am probably too punitive on the out years for margins. If I change my margin assumption to ~22% average EBITDA margins for 2024-2028, the value is ~$275BN. This is solid improvement, but seems well captured in Amazon’s $1.6 Trillion market cap.

2 thoughts on “Re-visting my AWS Thesis and Valuation $AMZN”

  1. I just landed up on your blog and I like the quality of your thinking. I just started looking into the cloud industry, and could it be that what’s potentially very high margin is not the infrastructure itself but the services that run on it? And could it be that your sales forecast just reflects the infrastructure component?

  2. That’s certainly possible. I guess my question is what services? And if there’s an endgame where maybe AWS becomes more valuable because AWS actually starts to offer SAAS services built off their own cloud… but does that mean they end up competing with their customers? Do they bite the hand the feeds them?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.