What is true is that Amazon managers are demanding substantially more work out of every employee. It has always been an intense place, but my former coworkers who have been at the company for a decade say that teams now have more responsibilities with fewer people. We see more frequent “large-scale events” — for example, AWS outages like the one that rocked the internet this October — because teams are under pressure to deliver more with less. Non-stop rumors of layoffs instill a corrosive fear into jobs we used to find challenging and satisfying. This fear means that the company can get away with unreasonable expectations of productivity.

Not only does Amazon demand we use AI internally, the company is rushing to build data centers because AWS sells the computing power that many other companies’ models run on. For example, Palantir runs the mass deportation software used by ICE on Amazon Web Services. Thus, layoffs can be understood both as an attempt to free up cash for more data centers and as a sales strategy, signaling to enterprise customers that they too will be able to lay off huge swaths of their corporate employees — all they have to do is sign up for AWS’s AI tools and services.