按年被提到
20252025 · 2 次2026
AI 安全、伦理与社会影响›AI 的经济与社会影响·arXiv:2508.15734 · cs.AI

Measuring the environmental impact of delivering AI at Google Scale

Cooper Elsworth, Keguo Huang, David Patterson, Ian Schneider, Robert Sedivy, Savannah Goodman, Ben Townsend, Parthasarathy Ranganathan, Jeff Dean, Amin Vahdat, Ben Gomes, James Manyika · 2025
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2025-08-27 · 最近 2025-09-15
2不同的人
2活跃月份
2讨论串
46被引(OpenAlex)
arXiv 摘要(候选层不写导读)

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.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论nerevarthelame · 2025-08

“> Didn't google just prove there is little to no environmental harm, INCLUDING if you account for training? I don't think they, have, no. Perhaps I'm overlooking something, but their most recent technical paper [0], published less than a week ago, states, "This study specifically considers the inference and serving energy consumption of an AI prompt. We …”

I Am An AI Hater原评论 ↗
评论Veedrac · 2025-09

“The energy use claims are questionable, but I at least get where they're coming from. The water use is the confusing part. Who looks at a server rack and goes ‘darn, look at how water intensive this is’? People use water as a coolant in large part because it's really hard to boil, plus it's typically cheap because it regularly gets delivered to your front…”

Gentoo AI Policy原评论 ↗