How does this company make money?
Amazon earns a commission each time a third-party seller makes a sale on its marketplace. AWS customers pay based on how much computing power and storage they actually use each month. Prime members pay a recurring subscription fee. Sellers and brands pay advertising fees to have their products appear at the top of search results. Amazon also sells products directly as a retailer itself.
What makes this company hard to replace?
FBA sellers already have their physical inventory sitting inside Amazon's fulfillment centers and their back-end systems connected to Amazon's warehouse management software — moving would mean physically retrieving that stock and rebuilding those integrations elsewhere. AWS customers who have built applications using Amazon-specific services like DynamoDB or Lambda would have to rewrite significant amounts of code before they could move to another cloud provider. Prime members are automatically billed for a 12-month subscription, which creates a year-long window before leaving is even a simple decision.
What limits this company?
During busy seasons like the holidays, the number of orders that can actually reach customers is capped by physical things — delivery vans, drivers, and local sorting facilities. Every extra Prime order needs a real vehicle and a real person at the door. When the volume of Prime-eligible orders exceeds what Amazon's own delivery fleet, its delivery service partners, and USPS can handle together, no amount of extra computing power fixes that. Unlike software, last-mile delivery cannot be scaled up from a server room.
What does this company depend on?
Amazon cannot operate without USPS, which handles rural and residential deliveries that Amazon's own fleet cannot reach cheaply. It also relies on Visa and Mastercard to process payments, UPS and FedEx as backup when internal delivery capacity runs out, the utility power grids supplying electricity to data centers across 33 regions worldwide, and third-party sellers who supply 60% of all units sold on the platform.
Who depends on this company?
Third-party marketplace sellers have their inventory physically sitting in Fulfillment by Amazon warehouses and their order systems wired into FBA's APIs — if that access ended, those sellers would be left with logistics arrangements they could not easily replace. AWS enterprise customers running live production workloads on EC2 instances would face application outages if they had to migrate away. Whole Foods stores rely on Amazon's supply chain systems to keep their shelves stocked.
How does this company scale?
AWS software services — storage, computing, databases — can be sold to one more customer at almost no extra cost once the data centers are built. That part scales very cheaply. Last-mile delivery does not work that way: every additional order in a new neighbourhood requires another vehicle, another driver, and another local sorting facility. That physical infrastructure has to grow one piece at a time, no matter how efficient the software behind it becomes.
What external forces can significantly affect this company?
GDPR and other data localization laws in various countries require AWS to keep certain customers' data inside specific borders, which means building and maintaining separate data center infrastructure in each jurisdiction rather than pooling it freely. U.S.-China trade tensions create uncertainty for cross-border fulfillment and could restrict where Amazon can offer cloud services. In rural areas, limited broadband speeds make it harder to deploy AWS edge computing for applications that need very fast response times.
Where is this company structurally vulnerable?
If the us-east-1 data center region in Virginia went down, Amazon's retail order processing and its outside enterprise cloud customers would fail at exactly the same moment — because both run on the same physical infrastructure. A competitor that kept its internal and external systems separate would only lose one revenue stream during an outage. Amazon's model, where the cost savings and the failure risk come from the same design choice, means one regional outage hits everything at once.
CompanyGraph