We always assumed our AWS bill was high because we were growing. Turns out… not really.

When we actually dug in, about 30% of our EC2 instances were sitting under 15% CPU for months. Staging was fully provisioned 24/7 even though it was only used maybe 10–15 hours a week. And no one touched any of it because ownership was unclear.

We had dashboards, but they didn’t answer the basic stuff:
what’s underutilized, who owns it, and what can safely be turned off.

On top of that, most of our spend was still on on-demand, while our Savings Plans didn’t really line up with how we were actually using resources.

After cleaning up idle stuff and fixing some of the commitment mismatch, we saw around a 20–25% drop in the bill within a couple weeks.

At this point it feels less like a scaling problem and more like a visibility + ownership problem.

Curious if others have seen something similar?