5. Inputs and Evidence Sources

The methodology draws on the following categories of inputs.

5.1 Hardware Benchmarking Data

Energy measurements collected from physical AI inference workloads running on real GPU hardware. GPU-board power is captured via the NVIDIA Management Library (NVML) at the configured polling interval, with non-GPU server energy estimated from wall time.

5.2 Model and Architecture Data

Publicly available and vendor-published information about AI model characteristics, including parameter counts, active parameter counts for mixture-of-experts architectures, quantisation levels, and GPU memory requirements. For open-source models, this information is typically available directly from model documentation and hosting platforms.

For proprietary and closed-source models where architecture details are not publicly disclosed, Greenpixie estimates model characteristics using a combination of third-party research, published performance benchmarks, architectural inference from comparable open-source models, and known architecture disclosures from earlier model generations. These estimates are cross-validated against economic plausibility and are subject to internal review. The estimation process is described in Section 6.4.

5.3 Cloud Provider Billing and Pricing Data

Publicly available pricing structures and billing configurations from AWS, Azure, and GCP, used to identify token-type distinctions, prompt-length tiers, batch processing options, and caching discounts.

5.4 Embodied Emissions Data

Hardware manufacturing and lifecycle emissions data adapted from the Boavizta methodology and associated formulae for server and GPU embodied emissions calculations.

5.5 Carbon Intensity Data

Regional electricity grid carbon intensity factors, as described in the core Greenpixie Methodology for Cloud Emission Measurement.

5.6 Published Research and Industry Sources

Publicly available research from institutions and organisations including the International Energy Agency, MIT Technology Review, Google, and others, used to inform contextual understanding and methodological design. A full bibliography is provided in Section 12.

5.7 Water Consumption Data

Cloud provider Water Usage Effectiveness values and regional water intensity factors, as set out in the Greenpixie Methodology for Cloud Emission Measurement.