With (most likely) 0 knowledge about the internal details of their business, but with the confidence of the average programmer: "they spend too much".

I imagine Twitter could also be built over a weekend, right? :-)

The more ignorant you are about the reality on the ground, apparently the easier it is for some people to tell others "obviously, you're doing it wrong".

Then again, maybe these savings come from whole same clean-up of services and other stuff they are running... After all there is likely decades of cruft around...

Literally every one of their vehicles is a knapsack problem and a traveling salesman problem needing to be solved every day. I'm sure they go through plenty of compute power.

With airplanes especially, optimal utilization is key. Routing millions of packages a day through such a complex network around weather and other incidents while keeping SLAs is simply astounding.

Yep... A couple of million of packages daily + tracking a fleet of vehicles, client facing site, apis for their tracking and courrier services/devices plus some internal services (billing etc...).. this seems like it would fit on a few racks, not nearly 400million, or whatever the full number is (since 400m are just the savings).

But yeah... old company, they might still have stuff written in cobol virtualized somewhere.

Their Memphis air hub processes 2 million packages by itself during the season. You also forgot to include FedEx Freight in your estimate.

edit - Autocorrect was terrible on my phone today!... The Fedex Memphis air hub processes more than 2 million packages a day itself during the winter holiday peak season.

With 300k employees you can expect them to have thousands of disparate services that will likely get rewritten part of such migration

But even virtualized Cobol should fit in a few racks.