Amazon Just Did the AWS Move, Again. This Time With Trucks.
Amazon Supply Chain Services turns a previously-internal cost (freight, fulfillment, last-mile) into an external revenue line. Same playbook as Marketplace and AWS. Can physical logistics compound the way compute did?
The interesting thing about Amazon Supply Chain Services is not that it launched. It is that it is the third time we are watching Amazon run the same play.
In 2000, Amazon took a fixed cost (its retail website, its catalog, its payments rails) and opened the doors to outside merchants. That became Amazon Marketplace. In 2002, Amazon began standardizing the internal compute infrastructure it had built for its own retail engine. By 2006, that effort had a public face: S3 in March, EC2 in August. That became AWS. On May 4 2026, Amazon opened its logistics network to every business, regardless of whether they sell on the marketplace. That is now Amazon Supply Chain Services. The press release quotes Peter Larsen, the VP running the new business, drawing the analogy to AWS himself.
In our view, the headline trade (AMZN to a 52-week high, UPS down roughly 10%, FedEx down roughly 7% on the May 4 print) is not the story worth thinking about. The story is structural. When a previously-internal cost migrates onto Amazon’s revenue line, every analyst model that treated logistics as a cost-of-goods problem now has to reframe it as a segment economics problem. That reframing has happened before, and the reader’s mental model for how it goes is the question worth working through.
The Pattern, Stated Plainly
The Bezos move has three repeatable features. First, latent infrastructure: capacity Amazon built for itself that ended up over-provisioned for retail alone. Second, a productization step (the seller portal in 2000, the AWS console in 2006, the new supplychain.amazon.com console in 2026) that turns a back-office system into a self-serve product. Third, marquee anchor customers that signal credibility: Netflix and Capital One were early AWS proof points; Procter & Gamble, 3M, Lands’ End, and American Eagle Outfitters are the launch tenants for Supply Chain Services.
The infrastructure being externalized is not modest. Amazon disclosed the asset base at launch: more than 80,000 trailers, more than 24,000 intermodal containers, an air fleet of more than 100 aircraft, and over 200 fulfillment centers in the United States. Functionally, Supply Chain Services is a re-bundling of pieces that already existed under separate names (Fulfillment by Amazon, Buy with Prime, Amazon Freight, Amazon Air Cargo, Supply Chain by Amazon for ocean), wrapped in a single console with AI-driven inventory placement and forecasting, as FreightWaves laid out in detail.
| Move | Year externalized | Revenue line today | Margin profile |
|---|---|---|---|
| Marketplace | 2000 | Third-party seller services, ~$172B in 2025 | High-margin commission stream |
| AWS | 2006 | AWS, ~$108B+ in 2024 | High-20s to 30%+ operating margin |
| Supply Chain Services | 2026 | Not yet a disclosed segment | TBD (see Section 5) |
If you are an analyst, the question this table forces you to answer is simple: which of those three rows does ASCS most resemble? That is what the rest of this piece tries to think through.
The Bull Case Borrows From AWS
The AWS arc is the cleanest precedent and the one Amazon explicitly invokes. Internal in 2002. Simple Queue Service publicly available in November 2004. S3 and EC2 in 2006. Roughly $21 million in revenue that first full year. By 2012, public estimates had AWS above $1.5 billion. By 2017, $17.5 billion. By 2020, north of $46 billion. AWS now anchors the largest profit pool inside Amazon and consistently runs operating margins in the high-20s to mid-30s, a structurally different business than the retail operation that birthed it.
The bull case for Supply Chain Services applies that math: take an over-provisioned physical network, productize it, attach AI on top, and let third-party demand absorb the spare capacity. The shape of the curve, if it works, looks similar: slow first, then compounding. We covered the broader question of how much AI capex the market is currently pricing in and the AI plumbing layer is part of why ASCS gets pitched as a software story rather than a freight story. The forecasting and inventory placement engine is software. The trailers and the fulfillment buildings are not.
The Bear Case Is Structural
There are three reasons the AWS analogy may not fully transfer, and a careful reader should weight each one.
The first is capex intensity. AWS scaled on servers and data centers, expensive, but riding falling unit costs from Moore’s Law and a commoditizing compute stack. Logistics scales on steel: trailers, planes, containers, distribution centers. Unit costs in physical capital do not deflate the way semiconductors do. Operating leverage exists, but the curve is shallower. Amazon’s 2024 capex was already roughly $78 billion; pushing the logistics half of that even harder takes years to amortize.
The second is the competitive structure. Cloud compute in 2006 was a green-field market with no public-company incumbents pricing aggressively. Parcel and freight is a mature market with credible incumbents (UPS, FedEx, DHL, DSV, Kuehne+Nagel) who have decades of customer relationships and will defend share with price. Citi analyst Ariel Rosa was direct: Amazon does not have the scale or physical network to displace all competitors. The closer Supply Chain Services gets to pure parcel, the more it begins to resemble a price war on a low-margin product, not a margin-expansion opportunity.
The third is conflict-of-interest. AWS’s customers were neutral: Netflix, Capital One, the public sector. They had no reason not to pay Amazon. Supply Chain Services’ addressable market includes Amazon’s competitors and suppliers in retail. Some, like P&G and 3M, will use it. Others, particularly large retailers and direct-to-consumer brands threatened by Amazon’s first-party catalog, will not, on principle. Walmart GoLocal and Shop Promise from Shopify exist precisely because some merchants do not want Amazon to own their fulfillment relationship. The addressable market for ASCS is structurally smaller than AWS’s was, and the reason is conflict that did not exist in compute.
The UPS Read-Through
The launch of Supply Chain Services is the other half of a story we have already covered from the UPS side. UPS pre-emptively chose to cut Amazon volume by more than 50% by June 2026, on a published timeline, and is targeting a 9.6% operating margin on roughly $89.7 billion in 2026 revenue, as the company reaffirmed on its Q1 2026 call. That decision looks materially different in light of May 4. UPS was not abandoning a profitable customer in a panic. UPS read the signal correctly: Amazon was always going to externalize the logistics network, and the strategic move was to fire Amazon as a customer before Amazon turned into a competitor.
Read together, the UPS glide-down and the ASCS launch are two ends of one story, not two separate stories. The carriers that lean into premium B2B, healthcare, cross-border, and high-touch freight will live somewhere different from the carriers that try to keep competing with Amazon on commodity parcel. Some of this dynamic also showed up in Amazon’s Q1 2026 print, where AWS growth carried the segment narrative and logistics was treated as cost discipline, and in the Graviton deal Amazon, Meta, and Intel announced earlier this year, where the same AWS playbook was running in custom silicon.
What To Watch in the Next Four Quarters
For a reader trying to monitor whether Supply Chain Services is becoming a real segment or staying a marketing rebrand, there are six items worth tracking.
The first is line-item disclosure. The “third-party seller services” line in Amazon’s segment reporting ran roughly $172 billion in 2025. ASCS revenue from non-marketplace customers will start to flow through that line first. The more interesting moment is when, or whether, Amazon breaks ASCS out as its own segment. AWS got that treatment in Q1 2015 after years inside “Other,” and the disclosure shift is what unlocked the rerating.
The second is capex run-rate. An incremental acceleration above the existing depreciation curve, specifically tied to logistics capacity rather than data centers, is the cleanest read on whether Amazon is leaning into ASCS demand or letting it absorb spare capacity.
The third is earnings-call language. Mentions of “Supply Chain Services,” “Amazon Freight,” and the AI forecasting platform in transcripts. AWS got its first dedicated earnings-call paragraph 18 to 24 months before its segment disclosure. Watch for the same arc.
The fourth is the customer roster. P&G, 3M, Lands’ End, and American Eagle are launch logos. The next six to twelve months of customer announcements will signal whether the addressable market is widening into new verticals or stalling out at retail-and-CPG.
The fifth is parcel-incumbent commentary. If UPS or FedEx flag “lost B2B accounts to Amazon” on their calls, that is direct ASCS share-take evidence in the wild.
The sixth is the counterforce. Major capability announcements from Walmart GoLocal or Shopify’s Shop Promise would signal that the merchant-conflict friction is real and is producing a parallel network.
How We Are Thinking About It
Our base case is that Supply Chain Services becomes a meaningful, disclosed revenue line within two to three years, and that the analyst community gradually re-segments AMZN to reflect it. Our skeptical case is that the margin profile settles closer to a low-double-digit operating margin (closer to a logistics business than to AWS) because of capex intensity and conflict-of-interest dynamics that did not exist in compute. The bull case is that the AI-driven forecasting and inventory placement engine creates enough software-style operating leverage to push margins meaningfully above incumbent carriers, in which case the AWS analog holds and the segment ends up much larger than current consensus expects.
We do not think readers should be making position changes today on the basis of a launch announcement and a one-day stock move. The interesting work is in the framework, not the trade. The right question to be asking is not whether to chase AMZN at a 52-week high or sell UPS into a 10% drop. The right question is whether your existing AMZN exposure, and your existing exposure to the carriers, the freight forwarders, and the retailers in your portfolio, already reflects a world where Amazon’s logistics network is a competitor to all three.
Watch the segment disclosure. Watch the capex line. Watch the customer roster. The signal is in the structure, not the headline.
Disclosure: Ferrante Capital LLC is a Registered Investment Adviser. This post discusses Amazon (AMZN), United Parcel Service (UPS), and FedEx (FDX). Ferrante Capital, its principals, and its clients may hold long or short positions in any of these securities at the time of writing or in the future. Holdings are subject to change without notice. Mention of any security is not a recommendation to buy, sell, or hold that security.
Forward-looking statements reflect Ferrante Capital’s current analysis and are based on information believed to be reliable at the time of writing. Forward-looking statements involve known and unknown risks, including changes in market conditions, regulatory developments, competitive dynamics, customer adoption rates, and macroeconomic factors. Actual outcomes may differ materially from any expectations expressed or implied in this analysis.
This article is for educational and informational purposes only and does not constitute investment advice. It does not take into account any individual reader’s financial situation, objectives, or risk tolerance. Please consult a qualified financial professional before making investment decisions.
