Chasing Carbon: The Elusive Environmental Footprint of Computing

By Udit Gupta, Young Geun Kim, Sylvia Lee, Jordan Tse, Hsien-Hsin S. Lee, Gu-Yeon Wei, David M. Brooks, Carole-Jean Wu. 2020

376 citations recorded by the source corpus.

Given recent algorithm, software, and hardware innovation, computing has enabled a plethora of new applications. As computing becomes increasingly ubiquitous, however, so does its environmental impact. This paper brings the issue to the attention of computer-systems researchers. Our analysis, built on industry-reported characterization, quantifies the environmental effects of computing in terms of carbon emissions. Broadly, carbon emissions have two sources: operational energy consumption, and hardware manufacturing and infrastructure. Although carbon emissions from the former are decreasing thanks to algorithmic, software, and hardware innovations that boost performance and power efficiency, the overall carbon footprint of computer systems continues to grow. This work quantifies the carbon output of computer systems to show that most emissions related to modern mobile and data-center equipment come from hardware manufacturing and infrastructure. We therefore outline future directions for minimizing the environmental impact of computing systems.

Subfields: Computer Science Applications, Electrical and Electronic Engineering, ML Systems & Hardware, AI Ethics, Fairness & Society, Human-Computer Interaction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Science Applications

In Electrical and Electronic Engineering

In ML Systems & Hardware

In AI Ethics, Fairness & Society

In Human-Computer Interaction

Citations in the corpus

Efficient Large Scale Language Modeling with Mixtures of Experts (2021, background): “There is also a concern about the environmental costs associated with training and deploying such models (Strubell et al., 2019; Gupta et al., 2021; Bender et al., 2021; Patterson et al., 2021) motivating research into more efficient model designs (Lepikhin et al.”
Carbon Emissions and Large Neural Network Training (2021, neutral): “Note that we are evaluating the CO 2 e of operating computers and datacenters, but not fabricating and recycling them (see [Gup20] for the latter topic).”
OPT: Open Pre-trained Transformer Language Models (2022, background): “We need to understand the manufacturing (or embodied) carbon of these systems (Gupta et al., 2021)”
Sustainable AI: Environmental Implications, Challenges and Opportunities (2021, background): “However, system design with a focus on operational energy efficiency optimization does not always produce the most environmentallysustainable solution (Jain & Wullert, 2002; Chang et al., 2010; Gupta et al., 2021).”
Evaluating the Social Impact of Generative AI Systems in Systems and Society (2023, background): “However, recent estimates suggest that the manufacturing process have substantial environmental costs [96].”
Measuring the Carbon Intensity of AI in Cloud Instances (2022, background): “Other studies have gone beyond simply tracking the emissions from training models, aiming to quantify the emissions resulting from manufacturing computing hardware [15], the broader impacts of sustainable AI [49], and the methodologies used to assess those impacts [21, 26].”
Socio-Technological Challenges and Opportunities: Paths Forward (2021, background): “…and computing technology sectors consume a significant amount of global electricity, water, and natural resources, leading to paramount carbon and environmental footprint [Jain and Wullert, 2002, Strubell et al., 2019, Manne, 2020, Wu and Gupta, 2021, Gupta et al., 2021, Crawford et al., 2021].”
Evaluating the carbon footprint of NLP methods: a survey and analysis of existing tools (2021, neutral): “Similarly, another recent study (Gupta et al., 2021) reports that the hardware manufacturing and infrastructure accounts for the bulk of the environmental impact of mobile and data center computing equipment, while the impact of operational energy consumption is diminishing.”
Unraveling the hidden environmental impacts of AI solutions for environment (2021, background): “In [14] the authors point out the methodological gaps of the previous studies focusing on the use phase.”
A Holistic Approach for Designing Carbon Aware Datacenters (2022, background): “However, from the capital perspective, over-provisioned servers increase the overall carbon footprint by increasing embodied carbon emissions from hardware manufacturing [22].”
ACT: designing sustainable computer systems with an architectural carbon modeling tool (2022, background): “More specifically, nearly half of Apple’s emissions from hardware manufacturing come from fabricating integrated circuits (IC’s) such as processors, memory, storage [8, 36].”