⚠ Overbuilt CapacityModerate threat

A warehouse web sized for endless growth is a liability the moment growth pauses, and Amazon has felt this before. During the pandemic it roughly doubled its fulfillment network on the belief that the surge in online shopping would hold; when demand normalized, it found itself with too much space and too many people, and spent the following years painfully working off the excess. The very scale that is a moat in good times is a fixed cost that cannot be shed quickly in bad ones.

The danger is structural: warehouses, leases, and equipment are long-lived commitments made against forecasts that can be wrong. Build too little and you lose the speed that defines you; build too much and you carry idle capacity through a downturn, watching fixed costs eat the thin retail margin. The faster Amazon grows the network, the larger the bill if the demand it assumed fails to arrive.

It helps that Amazon learned the 2022 lesson expensively and now manages capacity with far more discipline, using its data to add regional capacity closer to demand rather than in one great national build-out. Robotics and productivity gains also let the same footprint handle more volume¹, softening the need to overbuild.

The right rating is moderate — cyclical rather than structural. Overbuilding can produce an ugly year or two whenever growth disappoints, and the network's fixed costs make retail earnings swing more than the revenue does — but the capacity is not stranded, demand does eventually catch up to it, and Amazon has shown it can grow into an overbuilt network rather than being sunk by one.