# Measuring the environmental impact of delivering AI at Google Scale

arXiv: https://arxiv.org/abs/2508.15734 · HTML: https://mlsanity.com/paper/2508.15734/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 37 references that are in the map.
- First posted on arXiv 21 Aug 2025.

## Builds on (most cited first, 19 of 19)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference](https://mlsanity.com/paper/2403.04132/) (2024)
- [Carbon Emissions and Large Neural Network Training](https://mlsanity.com/paper/2104.10350/) (2021)
- [Estimating the Carbon Footprint of BLOOM, a 176B Parameter Language Model](https://mlsanity.com/paper/2211.02001/) (2022)
- [The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink](https://mlsanity.com/paper/2204.05149/) (2022)
- [From Words to Watts: Benchmarking the Energy Costs of Large Language Model Inference](https://mlsanity.com/paper/2310.03003/) (2023)
- [A Survey on Large Language Model Acceleration based on KV Cache Management](https://mlsanity.com/paper/2412.19442/) (2024)
- [Pathways: Asynchronous Distributed Dataflow for ML](https://mlsanity.com/paper/2203.12533/) (2022)
- [Zeus: Understanding and Optimizing GPU Energy Consumption of DNN Training](https://mlsanity.com/paper/2208.06102/) (2022)
- [Chasing Carbon: The Elusive Environmental Footprint of Computing](https://mlsanity.com/paper/2011.02839/) (2020)
- [MemServe: Context Caching for Disaggregated LLM Serving with Elastic Memory Pool](https://mlsanity.com/paper/2406.17565/) (2024)
- [Counting Carbon: A Survey of Factors Influencing the Emissions of Machine Learning](https://mlsanity.com/paper/2302.08476/) (2023)
- [The ML.ENERGY Benchmark: Toward Automated Inference Energy Measurement and Optimization](https://mlsanity.com/paper/2505.06371/) (2025)
- [Life-Cycle Emissions of AI Hardware: A Cradle-To-Grave Approach and Generational Trends](https://mlsanity.com/paper/2502.01671/) (2025)
- [Carbon accounting in the Cloud: a methodology for allocating emissions across data center users](https://mlsanity.com/paper/2406.09645/) (2024)
- [Misinformation by Omission: The Need for More Environmental Transparency in AI](https://mlsanity.com/paper/2506.15572/) (2025)

## Built on by (most cited first, 30 of 32)

- [A benchmark for vericoding: formally verified program synthesis](https://mlsanity.com/paper/2509.22908/) (2025)
- [Energy Use of AI Inference, Efficiency Pathways, and Test-Time Scaling](https://mlsanity.com/paper/2509.20241/) (2025)
- [Where Do the Joules Go? Diagnosing Inference Energy Consumption](https://mlsanity.com/paper/2601.22076/) (2026)
- [Tycho: Active Abstraction with Programmatic World Models for ARC-AGI-3](https://mlsanity.com/paper/2607.28287/) (2026)
- [AgentStop: Terminating Local AI Agents Early to Save Energy in Consumer Devices](https://mlsanity.com/paper/2605.15206/) (2026)
- [A Survey on the Green Development of Large Models: From Resource-Efficient Architectures to Hardware-Software Co-Design](https://mlsanity.com/paper/2607.09084/) (2026)
- [Towards a future space-based, highly scalable AI infrastructure system design](https://mlsanity.com/paper/2511.19468/) (2025)
- [VoltanaLLM: Energy-Efficient and SLO-Aware Disaggregated LLM Serving via Adaptive Frequency Control and State-Space Routing](https://mlsanity.com/paper/2509.04827/) (2025)
- [The Energy Footprint of LLM-Based Environmental Analysis: LLMs and Domain Products](https://mlsanity.com/paper/2604.00053/) (2026)
- [Neuromorphic Intelligence](https://mlsanity.com/paper/2509.11940/) (2025)
- [A Locally Executable AI System for Improving Preoperative Patient Communication: A Multi-Domain Clinical Evaluation](https://mlsanity.com/paper/2510.01671/) (2025)
- [Energy Scaling Laws for Diffusion Models: Quantifying Compute in Image Generation](https://mlsanity.com/paper/2511.17031/) (2025)
- [The environmental impact of ICT in the era of data and artificial intelligence](https://mlsanity.com/paper/2601.06174/) (2026)
- [Emissions and Performance Trade-off Between Small and Large Language Models](https://mlsanity.com/paper/2601.08844/) (2025)
- [Selecting Language Models for Social Science: Start Small, Start Open, and Validate](https://mlsanity.com/paper/2601.10926/) (2026)
- [Balancing Sustainability And Performance: The Role Of Small-Scale LLMs In Agentic Artificial Intelligence Systems](https://mlsanity.com/paper/2601.19311/) (2026)
- [AI Cap-and-Trade: Efficiency Incentives for Accessibility and Sustainability](https://mlsanity.com/paper/2601.19886/) (2026)
- [The Global Landscape of Environmental AI Regulation: From the Cost of Reasoning to a Right to Green AI](https://mlsanity.com/paper/2603.00068/) (2026)
- [Is the future of AI green? What can innovation diffusion models say about generative AI's environmental impact?](https://mlsanity.com/paper/2603.21419/) (2026)
- [LLM-generated personalized nudges for improving pro-environmental behavior: Field evidence from resource conservation](https://mlsanity.com/paper/2604.03881/) (2026)
- [Transparent Screening for LLM Inference and Training Impacts](https://mlsanity.com/paper/2604.19757/) (2026)
- [Scalable LLM-based Coding of Dialogue in Healthcare Simulation: Balancing Coding Performance, Processing Time, and Environmental Impact](https://mlsanity.com/paper/2604.23255/) (2026)
- [OpenG2G: A Simulation Platform for AI Datacenter-Grid Runtime Coordination](https://mlsanity.com/paper/2605.05519/) (2026)
- [Greening AI Inference with Accuracy and Latency-aware User Incentives](https://mlsanity.com/paper/2605.27309/) (2026)
- [Evaluation of ML Resource Utilization Requires Model Life Cycle Assessment](https://mlsanity.com/paper/2606.07632/) (2026)
- [WattLayer: Get Layers Right to Estimate Inference Energy of Neural Networks](https://mlsanity.com/paper/2606.27841/) (2026)
- [Lights, Camera, Carbon: Architectural Scaling Laws for Video Generation Energy Consumption](https://mlsanity.com/paper/2607.04553/) (2026)
- [Quantifying the Relationship Between Clinical Safety and Environmental Impact in Therapeutic LLMs](https://mlsanity.com/paper/2608.11830/) (2026)
- [From Caveman to Expert Analyst: Energy Consumption of Variable LLM Tasks](https://mlsanity.com/paper/2608.12350/) (2026)
- [AgentDecarbonizer: Carbon-Aware Execution for AI Agents](https://mlsanity.com/paper/2608.20566/) (2026)

## Abstract (arXiv)

The transformative power of AI is undeniable - but as user adoption accelerates, so does the need to understand and mitigate the environmental impact of AI serving. However, no studies have measured AI serving environmental metrics in a production environment. This paper addresses this gap by proposing and executing a comprehensive methodology for measuring the energy usage, carbon emissions, and water consumption of AI inference workloads in a large-scale, AI production environment. Our approach accounts for the full stack of AI serving infrastructure - including active AI accelerator power, host system energy, idle machine capacity, and data center energy overhead. Through detailed instrumentation of Google's AI infrastructure for serving the Gemini AI assistant, we find the median Gemini Apps text prompt consumes 0.24 Wh of energy - a figure substantially lower than many public estimates. We also show that Google's software efficiency efforts and clean energy procurement have driven a 33x reduction in energy consumption and a 44x reduction in carbon footprint for the median Gemini Apps text prompt over one year. We identify that the median Gemini Apps text prompt uses less energy than watching nine seconds of television (0.24 Wh) and consumes the equivalent of five drops of water (0.26 mL). While these impacts are low compared to other daily activities, reducing the environmental impact of AI serving continues to warrant important attention. Towards this objective, we propose that a comprehensive measurement of AI serving environmental metrics is critical for accurately comparing models, and to properly incentivize efficiency gains across the full AI serving stack.