How AWS estimates embodied emissions of IT hardware: The science and technology behind the latest customer carbon footprint methodology

This article explains AWS's methodology for calculating embodied carbon emissions of IT hardware in data centers, now included in the Customer Carbon Footprint Tool.

• Embodied emissions measure manufacturing carbon for servers, storage, networking equipment using life cycle assessment
• Process-based LCA methodology replaces spend-based calculations for more accurate component-level emissions
• Four calculation pathways: PLCA-Eng (highest fidelity), parametric PLCA, RCA-LCA (uses AI/LLMs), and EIO fallback
• Four-step workflow: data generation, aggregation, amortization over 6-year service life, allocation to customers
• Uses Amazon MWAA, Glue, SageMaker, Bedrock for serverless pipeline orchestration and ML
• Allocates emissions based on physical resource usage, not spend
• Partnering with Fraunhofer and Open Compute Project on industry standardization

AWS developed a rigorous, science-based framework to measure IT hardware carbon emissions across millions of components, enabling accurate customer carbon footprint reporting and infrastructure decarbonization investments.