Carbon-aware computing

Scheduling or shifting AI inference and training workloads to the times and regions where the electricity grid is cleanest, to cut the carbon emissions of the compute.

grounded in: Latest AI/tech trend theme: 'AI's energy and carbon footprint — the environmental cost of datacenters is emerging as a mainstream counter-narrative to the AI buildout' (HN front page, 2026-07-12).

Connected concepts

Inference economics, Adaptive inference serving, Accelerator depreciation, Distributed inference