Environment Variables The Week in Green Software: More New Hosts!
Feb 26, 2026
New co-hosts introduce backgrounds in green software and sustainability. The conversation tackles AI's growing energy and water demands and how to make those impacts tangible. They debate location trade-offs for data centers and the complexity and inefficiencies of Kubernetes. The group also explores enabled emissions, ethical concerns around public AI releases, and the need for transparency and disclosure.
AI Snips
Chapters
Transcript
Episode notes
Mixed Units Hide Scale And Incentives
- Media articles use mixed units for water and energy metrics which obscures understanding of scale and intent.
- Asim found cubic kilometers, million liters, US gallons and trillions of liters in one article and traced a sponsor (Xylem), showing unit chaos can hide framing and incentives.
Normalize Sustainability Metrics For Clarity
- Standardize and normalize sustainability metrics to aid comparison and accountability.
- Carlos recommended simple normalization (e.g., relatable volumes like Titanic or Burj Khalifa equivalents) to avoid confusion across business units and articles.
Data Center Location Is A Multivariate Tradeoff
- Location choice for data centers requires multivariate tradeoffs between water, carbon, geology and energy supply.
- Oli explained Phoenix has low water but abundant baseload energy and solar, so placing a data center there reduces carbon but raises water impact.