Amazon: The Pain is Acute, Not Chronic
Amazon's revenue growth is something to behold:

Image upload

Amazon has seven distinct reporting segments. In order of revenue:
  1. Online Stores
  2. Third Party Seller Services
  3. AWS
  4. Ads
  5. Subscription Services
  6. Physical Stores
  7. Other

Currently segments 1,5,6 & 7 are losing money. Collectively these segments accounted for 53.9% of revenue last quarter:

Image upload

The unprofitable segments accounted for 63.83% of revenue in 2019.

The long Amazon thesis has two components:

  1. The more profitable segments are rapidly gaining share (driven by AWS and ads). The best time to buy a business is when it's at trough margins heading into a rapidly increasing margin cycle.
  2. Amazon is being held down by the "acute" problem of inflation, specifically higher transport costs (esp. diesel). This a short term problem, and will make Amazon stronger on a relative basis long term because of scale efficiencies. Greener pastures should be visible within a year which will lead to a re-rating of the business.

I'll address #1 first. Looking forward to 2024, the revenue mix should look something like the following:

Image upload
Note that AWS share of revenue is expected to increase by 41% (18.48% / 13.1% = 1.41).

Long term operating margins for the three profitable segments should be:

AWS: 30-35% (Amazon's own forecast)
Ads: 35%
3PL+: 13% (I've estimated this by looking at a breakdown of the business, assuming it's a combination of UPS + eBay style market place + warehousing/traditional 3pl)

The implied operating income of the three segments by 2024 would be:
AWS: $42B
Ads: $18.6B
3PL+: $20.6B

Collectively: $81.2B

One year from now this could be the forward looking operating income expectation for these segments. Using a reasonable 22X Multiple (Microsoft's current OI multiple in a bear market) brings us to a $1.79T market cap, representing 42% upside.

As for #2. Amazon's negative FCF is due to surging fulfillment costs. These costs are mostly fuel, but also the operating expenses related to having doubled their fulfillment capacity in less than 2 years.

This analysis comes from Morgan Stanley and was done earlier this year. I've updated the calculation to compare fulfillment costs from 2021 w/ diesel at $3.29 to when diesel hit its peak at $5.30 (which equated to fulfillment cost per unit of $4.66)

Image upload

2021: $3.29 X 21,095 = $69B
2022: $4.66 X 21,095 = $98B

So Amazon took a near $30B hit on fulfillment in a single year.

Image upload

Diesel has been coming down steadily and currently hovers around $5.00.

My estimate is around 45% of Amazon's fulfillment costs are related to Prime members and are hence Amazon's liability. For perspective, Amazon's entire subscriptions revenue in 2021 was only $31.8B.

Fulfillment costs should drop materially over the next year for three reasons:

  1. It reduced headcount by 100,000 last quarter alone. It over-hired expecting the pandemic related growth to continue. These costs will drop as they let attrition reduce their workforce.
  2. Amazon has 100,000+ diesel vans on the road today. Amazon has ordered 100,000 EVs from Rivian and more from Stellantis. EVs are expected to be delivering in over 100 cities by the end of this year. What's more important than the immediate fuel savings is that by the end of next year investors will have insight into how much savings we can expect from using EVs instead of diesel. Fuel-equivalent cost per mile using EV charging costs as a proxy imply a drop of at least half.
  3. Amazon launched "Buy with Prime". This allows sellers - even if they aren't selling on Amazon - to use Amazon's fulfillment, delivery, and return services. This will go a long way to helping Amazon amortize the investment into its fulfillment network across more deliveries & will bring unit costs down through scale efficiencies.

  1. Amazon has a world class moat built around AWS, it's fulfillment network (larger than UPS), Prime, network effects from online marketplace, and its brand.
  2. Unprofitable segments will have a clear path to break-even by the end of next year.
  3. We're looking at a 42% return if Amazon trades at 22X normalized operating income on a forward basis a/o year end 2023
  4. If the market recovers and Amazon is able to command a 25X normalized operating income it would be a $2T stock, representing 60% upside.

This is an exceedingly rare opportunity given Amazon's moat and the size of the business.
Nathan Worden's avatar
Unrelated side note— I was reading your piece on solving aging this weekend and thought it was super interesting and thought provoking. I know it wasn’t specifically about investing, but still had my mind mulling over all the implications. Was a fun piece. :)
Benjamin Buchanan's avatar
@nathanworden Really appreciate that Nathan! I have been talking to people about aging for many years, and thought it would be nice to have some of the big talking points that always come up in a single location.
Nathan Worden's avatar
@01core_ben it’s a timeless topic 😉
Conor Mac's avatar
@nathanworden +1, that was a fantastic piece. This is also the kind of stuff I enjoy reading outside of investing content too. Health, nutrition, et al.
Yegor's avatar
I hope me canceling my prime membership this year doesn’t impact your thesis (half jokingly) I’m actually experimenting to see how likely will I be to re sign up back aka if I’m okay with waiting more then 2 days to get my stuff or if I have been trained to get what I want asap
Benjamin Buchanan's avatar
@from100kto1m The funny thing about Prime is that Amazon loses money on it. It's key to the long term MOAT, but some attrition now wouldn't be the worst thing in the world for cash flow, sadly, lol!
Benjamin Buchanan's avatar
@01core_ben And I say this even taking into account the likely Ad revenue they make through Prime members - which is probably where the profits will come from longer term. Given their ridiculous content budget / fulfillment costs I'll be happy if Subs is breaking even by end of 24
Nathan Worden's avatar
Fantastic analysis!
Question: Would this thesis be broken if in a year transportation costs were the same or more than they are today?
Benjamin Buchanan's avatar
@nathanworden It would be elongated. I.e. the expectation about when they would hit the price target would change until that operating income was actually likely. Either way there's a BIG margin for error at these prices.
Nathan Worden's avatar
@01core_ben good point and point taken :)
Benjamin Buchanan's avatar
Let's gooooo!
Dissecting the Markets's avatar
Do you see the rise of the multicloud reducing the moat that AWS has built? The reason I say this is that a multicloud reduces the pricing power that these cloud computing platforms have had for the past several years.
Benjamin Buchanan's avatar
@dissectmarkets Rise of multi-cloud as in people using more than one of AWS/Azure or the intro of DDOG/Snowflake/etc other businesses
Benjamin Buchanan's avatar
@dissectmarkets Saw this on twitter this morning:
Image upload
Dissecting the Markets's avatar
@01core_ben Thanks for sharing this with me
Conor Mac's avatar
This is a stunning memo, well done for fitting that into 750 words. Feel like Amazon is written off so casually these days, but it's still got a long road ahead to become an even larger presence in our lives (it already is massively).
Benjamin Buchanan's avatar
@investmenttalk Thank you Conor!
Benjamin Buchanan's avatar
Just wrote this post with some thoughts on Cloud moats since AWS is such a big piece of the Amazon thesis, linking here:
Sagar Vensi's avatar
Amazon's marketplace business is so underrated because people tend to view it as one whole segment but it's not true because the 3P business is extremely profitable.