As CO2 emissions from data centers begin to overtake those of aviation, it’s time for software engineers to take responsibility for the carbon impact of the code we ship.

The current AI boom shows how finite energy and data center space is. In a constrained market efficient software means more than saving energy: It might decide if you can deploy at all.

But before we can optimize, we need to measure. How do we even quantify software efficiency? In this talk, we will introduce practical tools for measuring energy use and efficiency, work through examples of how to interpret the results, and show how these measurements can become part of your engineering feedback loop—helping you detect regressions and prevent systems from quietly becoming less efficient over time.

Java or Rust? Microservice or monolith? VM or bare metal? Or maybe serverless? Everything has impact and not always in the way you think.

You will leave with a practical framework for measuring software efficiency, comparing technical choices, and making better-informed decisions about the environmental and operational impact of the software you ship.