Pith. sign in

REVIEW 3 major objections 6 minor 4 cited by

Health-Informed Computing: Estimating and Addressing the Public Health Impact of Data Centers

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read U.S. data centers could impose $20.9 billion in public health costs by 2028.

desk verdict First credible COBRA-based screen of U.S. data center health burden; headline $20.9B rests on an untested average-attribution assumption, but the core environmental-justice and carbon-health decoupling results hold up. read the letter →

arxiv 2412.06288 v4 pith:ZJQE6ACA submitted 2024-12-09 cs.CY

Yuelin Han , Zhifeng Wu , Pengfei Li , Adam Wierman , Shaolei Ren This is my paper · ORCID
classification cs.CY
keywords datacenterspublichealthcriteriaairpollutantsAIsustainabilityCOBRAmodelhealth-informedcomputinggeographicalloadbalancingenvironmentaljustice
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper aims to quantify the hidden public health toll of U.S. data centers, which it frames as an overlooked consequence of AI's growth. Using EPA's COBRA model and LBNL electricity projections, it estimates that annual data-center-related health costs will rise from $6.7 billion in 2023 to $11.7–$20.9 billion in 2028, rivaling California's on-road vehicle emissions. The authors further claim that these costs fall unevenly, with per-household burdens in the most affected counties roughly 200 times higher than in the least affected, and that the worst-hit places are often not where data centers sit. They then propose health-informed computing, a load-scheduling framework that treats health cost per megawatt-hour as an objective and demonstrates a ~26% health-cost reduction without sacrificing electricity or carbon savings. The stakes are practical: if the estimates hold, data center growth is creating a multi-billion-dollar public health externality that current carbon-centric sustainability metrics do not capture.

What carries the argument

The argument rests on two tools. COBRA, the EPA's reduced-complexity Co-Benefits Risk Assessment model, couples the PCAPS air-dispersion source-receptor matrix with concentration-response functions, converting pollutant changes into county-level counts of deaths, asthma symptoms, and dollar damages; the paper feeds it with emissions scaled from baseline data. The second is the AVERT regionalization: the U.S. grid is split into 14 regions, and the paper assigns data center electricity to each region from EPRI state-level data, then scales every plant's baseline emissions down by the ratio $x\% = e_{DC}/e_{\mathrm{total}}$ in that region, exploiting COBRA's near-linearity. For the scheduling part, hourly marginal health price per megawatt-hour (from the data source cited as [77]) becomes the objective coefficient in a linear program that shifts workloads across 13 data center locations over 8,760 hours.

What would settle it

One decisive test is a bottom-up marginal-accounting comparison: take the same 2023 and 2028 electricity totals for U.S. data centers, attribute them to the actual marginal plants using plant-level dispatch data, and run COBRA on those emission changes; if the resulting national annual health cost differs from $6.7 billion or $20.9 billion by more than about 30%, the average-attribution method fails. A second observable check is whether region-level power plant emissions actually track data center load hour-by-hour; if they do not, the $x\% = e_{DC}/e_{\mathrm{total}}$ scaling assumption is unsupported.

Watch

Extended reading notes

Core claim

The central claim is that data centers, through their backup generators (scope 1) and the power plants that supply them (scope 2), will cause approximately 1,300 premature deaths and 600,000 asthma symptom cases per year in the contiguous U.S. by 2028, with a public health cost of $11.7–$20.9 billion depending on AI growth. The paper asserts that this burden is location-dependent and largely decoupled from carbon emissions: the spatial correlation between average health price and carbon intensity across 114 U.S. regions is only 0.292, and the per-household county-level ratio between the highest and lowest affected communities is about 200. It further shows that training a single Llama-3.1-scale model can produce air pollution equivalent to more than 10,000 LA-NYC car round trips and a health cost exceeding 120% of the training electricity cost. Finally, it claims that a health-informed geographical load balancing formulation, using marginal health price signals, cuts health cost by 25.8% relative to the 2023 baseline while simultaneously reducing electricity cost by 3.0% and carbon by 1.4%, and that carbon-aware load balancing can actually increase health costs.

Load-bearing premise

The biggest assumption is that data center electricity is drawn from the average mix of all power plants in each region, so the paper scales every plant's emissions down by the data center's share of regional electricity; if data centers actually pull from the dirtiest plants, the $20.9 billion figure and its county distribution would change.

Editorial extensions

If this is right

  • U.S. data center health costs are projected to roughly triple from $6.7 billion in 2023 to $11.7–$20.9 billion in 2028, making the industry's health burden comparable to California's on-road vehicle emissions.
  • The county-level per-household burden varies by about 200-fold, and every one of the top-10 most affected counties has a median income below the national median.
  • Training one Llama-3.1-scale model (about 30 GWh) can generate health costs of $0.23–$2.5 million depending on site, exceeding 120% of the training electricity bill in some locations.
  • A health-informed geographical load balancing algorithm cuts health cost by ~26% under modest slack ($\lambda = 1.5$) while also lowering energy cost and carbon, whereas carbon-aware scheduling alone can increase health cost.
  • Virginia backup generators at 10% of permitted emissions already cause roughly 14,000 asthma symptom cases and $220–$300 million in annual health costs across the region.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the appendix's scope-3 semiconductor facility health costs are scaled to AI chip demand, the paper's main totals would likely rise beyond $20.9 billion, since the single example facility contributes $26–39 million per year.
  • The average-attribution assumption may understate the marginal burden: if data centers are served by marginal gas or coal plants, the projected $20.9 billion is more likely a lower bound, and the county-level distribution would shift toward coal-belt states.
  • Because the spatial correlation between health price and carbon intensity is only 0.292, a health metric should be reported alongside carbon in sustainability disclosures; carbon-only regulation would miss the communities that carry the highest health burden.
  • One could validate the COBRA-based figures by running a full photochemical grid model on a few representative power-plant configurations; large discrepancies in downwind populated counties would indicate the reduced-complexity model needs recalibration for this application.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This paper quantifies the public health burden of criteria air pollutant emissions from U.S. data centers by decomposing emissions into scope 1 (onsite backup generators), scope 2 (electricity generation), and scope 3 (supply chain), and applying the EPA's COBRA reduced-complexity model to estimate health outcomes and monetized costs. Using LBNL/EPRI electricity projections, the authors estimate a total annual public health cost of $6.7 billion in 2023, rising to $11.7–$20.9 billion in 2028 under low/high growth scenarios, with scope 2 contributing roughly 92% of the high-growth total. They further report large county-level disparities, per-household burdens up to roughly 200 times higher in some low-income counties, location-dependent costs for training a Llama-3.1-scale model, and a health-informed geographical load balancing (HI-GLB) framework that achieves a claimed ~26% health cost reduction relative to a carbon-aware baseline.

Significance. If the central estimates are accepted, this is a timely and policy-relevant quantification of an overlooked externality of AI infrastructure. The paper's strengths include the use of EPA's validated COBRA tool, transparent reporting of low/high ranges, explicit acknowledgement of excluded components (cooling towers, scope 3), and a clear methodological pipeline from emissions to health costs. The proposed health-informed computing framework is a useful conceptual contribution, and the comparison with carbon-aware scheduling underscores an important distinction. The significance is tempered, however, by the fact that the headline national and county-level numbers rest on attribution assumptions that are plausible but not yet stress-tested.

major comments (3)
  1. [Appendix A.2, Section 5.1.3] The dominant scope-2 estimate is computed by scaling every power plant's baseline emissions in each AVERT region by x% = e_DC/e_total and applying COBRA linearly. This is an average-attribution accounting choice, but the resulting headline figures inherit two weakly tested premises: (a) the EPRI 2023 state-level data center load distribution, scaled to LBNL national totals, is fixed for all years 2019–2028; and (b) data centers are served by the regional average plant mix rather than the marginal plants that actually respond to new load. For the 2028 high-growth scenario, roughly 400 TWh of incremental demand is added beyond 2023, so the location and marginal generation mix of that incremental load, not the historical footprint, largely determine the regional x% values and thus the $20.9 billion total and the county-level map. The paper itself cites delayed coal retirements and new gas plants in Section 3.2 as consequences of data center growth, which suggests that marginal generation may be dirtier than the fleet average, while PPAs and dedicated renewables could push in the opposite direction. No sensitivity analysis or bounding scenario is provided for either premise. Because scope 2 accounts for about 92% of the 2028 high-growth estimate, this is a load-bearing uncertainty that needs to be addressed before the headline claim can be considered robust.
  2. [Section 6.2, Table 4] The HI-GLB evaluation uses the same health price metric as the optimization objective, so the reported ~26% reduction in health cost is partly a property of the objective rather than an independent validation of the framework. The health price from WattTime is a marginal, mortality-only signal, while the baseline national estimates use COBRA's average attribution with multiple health endpoints. To support the claim that health-informed scheduling 'can effectively mitigate health impacts,' the authors should validate the optimized schedules against an independent health metric (for example, COBRA-based average-attribution health costs computed from the resulting load distribution, or a different exposure-response function), or explicitly reframe the result as an in-sample optimized bound. As written, the comparison to carbon-aware GLB conflates objective-aligned cost reduction with demonstrated real-world health improvement.
  3. [Appendix A.1, Section 5.1] The national scope-1 estimate extrapolates the emission rate (tons/MWh) derived from Virginia's backup generator permits to all other states using data center electricity consumption. Generator emissions depend on permit emission limits, generator Tier composition, operating hours, and demand-response activation patterns, all of which vary substantially by state and utility. Virginia's unusually large and concentrated generator fleet may not be representative. The paper notes that scope-2 costs dominate, but scope-1 still contributes about $1.6 billion in the 2028 high scenario (Table 1), which is not negligible. A sensitivity analysis using alternative utilization assumptions or per-state generator data, or at least a clear statement of the resulting uncertainty, would strengthen the estimate.
minor comments (6)
  1. [Abstract and Section 1] The abstract and introduction emphasize 'lifecycle pollutant emissions' and scope 3, but the main empirical analysis excludes scope-3 impacts except for one illustrative facility in Appendix A.3. This should be stated more prominently in the abstract to avoid overstatement of the analyzed scope.
  2. [Table 1] The row labeled '2019 to 2023' reports cumulative five-year totals but appears under the 'Year' column alongside annual rows. It should be explicitly labeled as a cumulative period, e.g., '2019–2023 (cumulative)', to avoid confusion.
  3. [Figure 4] The legend entries 'Data Center (Low)' and 'Data Center (High)' refer to LBNL growth scenarios, while the paper also uses 'low' and 'high' for COBRA's health-cost bounds. The figure and Table 1 would benefit from a consistent naming convention, such as 'growth-low' and 'growth-high' for scenarios and 'estimate-low/estimate-high' for COBRA bounds.
  4. [Section 6.1, footnote 7] The claim that health costs exhibit greater temporal variation than carbon intensity in 110 out of 114 regions is based on WattTime's marginal health price, which considers only PM2.5 mortality. This important limitation is relegated to a footnote; it should be stated in the main text where the the comparison is made.
  5. [Appendix A, electricity price paragraph] The text says 'when estimating the electricity cost for data centers in 2023 and 2038,' but the paper analyzes 2028. This appears to be a typo.
  6. [Introduction, last paragraph of Section 1] Minor typo: 'wile AI and data centers offer many societal benefits' should read 'while AI and data centers offer many societal benefits.'
Assumptions & free parameters 1 free parameters · 6 assumptions · 0 invented entities

The central estimates rely on EPA COBRA as an external benchmark, plus several domain assumptions about attribution. One free parameter (10% generator utilization) is taken from state reports with a stated linear sensitivity. No new physical entities are postulated; 'health-informed computing' is an optimization criterion, not an entity.

free parameters (1)
  • Backup generator utilization rate = 10% of permitted emissions
    Section 3.1 and Appendix A.1 set actual generator emissions at 10% of permitted levels based on Virginia JLARC (~7%), Quincy WA (3-12%), and EPA demand-response allowances; the authors note results scale linearly with this percentage.
assumptions (6)
  • domain assumption COBRA v5.1 with PCAPS correctly estimates county-level PM2.5/ozone health impacts from power plant and generator emission changes.
    The paper uses COBRA as its dispersion and health model (Sections 4.2, 4.4, Appendix A) and does not independently validate it for data-center-specific emission profiles.
  • domain assumption Average emission attribution is the correct basis for data center health costs.
    Section 4.1.2 selects the average attribution method following carbon accounting practice; the authors acknowledge marginal rates exist but do not use them for the main estimates.
  • domain assumption Health impact scales linearly with proportional emission reduction within each AVERT region.
    Appendix A.2 applies x% emission reductions to all plants in a region based on COBRA's near-linearity; departure from linearity would distort county-level results.
  • domain assumption Virginia's scope-1 per-MWh emission rate applies to data centers in other states.
    Appendix A.1 multiplies Virginia's derived rate by other states' data center energy consumption because permit data elsewhere is unavailable.
  • domain assumption EPRI's 2023 state-level data center electricity distribution scaled to LBNL totals remains valid for 2019-2028.
    Appendix A.2 and A.6 use EPRI state shares, scaled by LBNL national totals, to compute regional x% values for all years.
  • domain assumption WattTime marginal health price is representative of health damage for load shifting.
    Section 6.1 and Appendix B.1 use WattTime health prices, which only include PM2.5 mortality, not the full set of COBRA endpoints.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Health-Informed Computing: Estimating and Addressing the Public Health Impact of Data Centers." pith.science (2026). https://pith.science/paper/ZJQE6ACA

@misc{pith2026241206288,
  author       = {Pith},
  title        = {Pith review of: Health-Informed Computing: Estimating and Addressing the Public Health Impact of Data Centers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZJQE6ACA}},
  note         = {Machine review of arXiv:2412.06288}
}
abstract

The surging demand for artificial intelligence (AI) has led to a rapid expansion of energy-intensive data centers, contributing to criteria air pollutant emissions and raising public health concerns that have received comparatively limited attention in sustainability assessments. This paper introduces a principled methodology to model air pollutant emissions for data centers and estimate the public health impacts. Our findings reveal that the growing demand for AI and computing technologies is projected to push the total annual public health burden of U.S. data centers up to more than $20 billion in 2028. Although national-level impacts remain modest, data center health costs are unevenly distributed: in the most affected counties, the estimated per-household health burden can reach about seven times the national average. Next, we propose a health-informed computing framework that explicitly incorporates public health impacts into data center resource management across space and time, mitigating public health costs while supporting environmental sustainability. More broadly, we recommend extended energy reporting to include public health impact of data centers and paying attention to all impacted communities.

Figures

Figures reproduced from arXiv: 2412.06288 by the authors.

Figure 1
Figure 1. The overview of data centers’ contribution to air pollutants and public health impacts. Scope-1 and scope-2 impacts occur during the operation of data centers (“operational”), whereas scope-3 impacts arise from activities across the supply chain (“embodied”). Under the Clean Air Act, the U.S. EPA is authorized to regulate the emission levels of criteria air pollu￾tants, reducing concentrations to comply with the Nat… view at source ↗
Figure 2
Figure 2. The county-level total scope-1 health cost of data center backup generators operated in Virginia (mostly in Loudoun County, Fairfax County, and Prince William County) [62]. The backup generators are assumed to emit air pollutants at 10% of the permitted levels per year. The total annual public health cost is $220-300 million, including $190-260 million incurred in Virginia, West Virginia, Maryland, Pennsylvania, New… view at source ↗
Figure 3
Figure 3. Public health costs of electricity generation and on-road emissions in the contiguous U.S. in 2023 and 2028 [39]. The error bars represent high and low esti￾mates returned by COBRA using two different exposure￾response functions. Based on the emission data projected by the U.S. EPA’s COBRA modeling tool [39], we show in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4 [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: The county-level total health cost of U.S. data centers from 2019 to 2023. (a) Health cost map; (b) CDF of county-level health cost; (c) Top-10 counties by total health cost. 5.1.2 Uneven distribution of data centers’ public health impacts. Next, [PITH_FULL_IMAGE:figu…
Figure 6
Figure 6. Figure 6: The county-level per-household health cost of U.S. data centers from 2019 to 2023. (a) Per-household health cost map; (b) CDF of county-level per-household health cost; (c) Top-10 counties by per-household health cost. IR represents “County-to-Nation Per-Household Medi…
Figure 7
Figure 7. Figure 7: The county-level per-household health cost of two U.S. technology companies in 2023. 5.2 Public Health Impact of Generative AI Training We now study the health impact of a specific computing task. Specifically, we consider the training of an LLM and assume that the ele…
Figure 8
Figure 8. Figure 8: Analysis of marginal scope-2 carbon emission rates and public health costs over 114 U.S. regions between October 1, 2023 and September 30, 2024 [77]. (a) In 110 out of the 114 U.S. regions (96%), the normalized IQR of marginal health cost is higher than that of margina…
Figure 9
Figure 9. Figure 9: Correlation analysis. (a) CDF of correlation coefficients between hourly health prices and marginal carbon emission rates for all the U.S. regions; (b) Scatter plot of health price and marginal carbon emission rate (annual average in 2023) across Meta’s U.S. data cente…
Figure 10
Figure 10. Figure 10: State-level electricity consumption of U.S. data centers in 2023 [5]. B Additional Results for Health-Informed GLB B.1 Details of the experiment setup We use Meta’s electricity consumption for each U.S. data center location in 2023 [37] for our experiments [PITH_FULL…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 9 citations worldwide. Full citation record

  1. CarbonFlex: Enabling Carbon-aware Provisioning and Scheduling for Cloud Clusters

    cs.DC 2025-05 conditional novelty 6.5 of 10

    A learning-based cluster scheduler that jointly provisions capacity and elastically scales batch jobs to cut operational carbon by up to 57.5% in AWS tests.

  2. Misinformation by Omission: The Need for More Environmental Transparency in AI

    cs.CY 2025-06 conditional novelty 6.0 of 10

    Environmental disclosure for notable AI models peaked in 2022 and then declined, and out-of-context energy and emissions estimates now dominate media coverage.

  3. The Cloud Next Door: Investigating the Environmental and Socioeconomic Strain of Datacenters on Local Communities

    cs.DC 2025-06 conditional novelty 5.0 of 10

    A preliminary mixed-methods case study argues that datacenters impose localized noise, power-quality, and economic burdens on neighboring communities and offers an equal-weight qualitative and quantitative framework f...

  4. AI Data Centers and the Water Use Feedback Loop

    cs.CY 2026-06 unverdicted novelty 3.0 of 10

    The paper formalizes the Water and AI Feedback Loop, introduces the Water Consumption Impact index, and shows water burden from AI data centers varies from 0.2% to 134% of local capacity across ten US sites.

Reference graph

Works this paper leans on

109 extracted references · 1 canonical work pages · cited by 4 Pith papers

  1. [1]

    Centers for Disease Control and Prevention

    U.S. Centers for Disease Control and Prevention. Artificial intelligence and machine learning: Apply- ing advanced tools for public health.https://www.cdc.gov/surveillance/data-modernization/ technologies/ai-ml.html

  2. [2]

    Mary Tran. U.S. Department of State DipNote: New air quality dash- board uses AI to forecast pollution levels.https://www.state.gov/ new-air-quality-dashboard-uses-ai-to-forecast-pollution-levels/, May 2024

    2024

  3. [3]

    How artificial intelligence and ma- chine learning can help healthcare systems respond to COVID-19.Machine Learning, 110:1–14, 2021

    Mihaela Van der Schaar, Ahmed M Alaa, Andres Floto, Alexander Gimson, Stefan Scholtes, Angela Wood, Eoin McKinney, Daniel Jarrett, Pietro Lio, and Ari Ercole. How artificial intelligence and ma- chine learning can help healthcare systems respond to COVID-19.Machine Learning, 110:1–14, 2021

    2021

  4. [4]

    Smith, Alex Hubbard, Alex Newkirk, Nuoa Lei, Md Abu Bakar Siddik, Billie Holecek, Jonathan Koomey, Eric Masanet, and Dale Sartor

    Arman Shehabi, Sarah J. Smith, Alex Hubbard, Alex Newkirk, Nuoa Lei, Md Abu Bakar Siddik, Billie Holecek, Jonathan Koomey, Eric Masanet, and Dale Sartor. 2024 United States data center energy usage report.Lawrence Berkeley National Laboratory LBNL-2001637, December 2024

    2024

  5. [5]

    EPRI. Powering intelligence: Analyzing artificial intelligence and data center energy consump- tion.White Paper on Technology Innovation Report, 2024.https://www.epri.com/research/products/ 3002028905

    2024

  6. [6]

    Department of Energy

    U.S. Department of Energy. Recommendations on powering artificial intelligence and data center infrastructure, Jul. 2024

    2024

  7. [7]

    Sustainable AI: Environmental implica- tions, challenges and opportunities

    Carole-Jean Wu, Ramya Raghavendra, Udit Gupta, Bilge Acun, Newsha Ardalani, Kiwan Maeng, Gloria Chang, Fiona Aga, Jinshi Huang, Charles Bai, et al. Sustainable AI: Environmental implica- tions, challenges and opportunities. InProceedings of Machine Learning and Systems, volume 4, pages 795–813, 2022. 19

    2022

  8. [8]

    Lee, Gu-Yeon Wei, David Brooks, and Carole-Jean Wu

    Udit Gupta, Young Geun Kim, Sylvia Lee, Jordan Tse, Hsien-Hsin S. Lee, Gu-Yeon Wei, David Brooks, and Carole-Jean Wu. Chasing carbon: The elusive environmental footprint of computing.IEEE Micro, 42(4):37–47, jul 2022

    2022

Show all 109 references
  1. [9]

    Islam, and Shaolei Ren

    Pengfei Li, Jianyi Yang, Mohammad A. Islam, and Shaolei Ren. Making AI less ’thirsty’.Commun. ACM, 68(7):54–61, June 2025

    2025

  2. [10]

    Environmental report.https://sustainability.google/reports/ google-2024-environmental-report/, 2024

    Google. Environmental report.https://sustainability.google/reports/ google-2024-environmental-report/, 2024

    2024

  3. [11]

    DCFlex initiative.https://msites.epri.com/dcflex, 2024

    EPRI. DCFlex initiative.https://msites.epri.com/dcflex, 2024

    2024

  4. [12]

    Deepspeed-MOE: Advancing mixture-of-experts inference and training to power next-generation AI scale

    Samyam Rajbhandari, Conglong Li, Zhewei Yao, Minjia Zhang, Reza Yazdani Aminabadi, Am- mar Ahmad Awan, Jeff Rasley, and Yuxiong He. Deepspeed-MOE: Advancing mixture-of-experts inference and training to power next-generation AI scale. InICML, 2022

    2022

  5. [13]

    Junkyard computing: Repurpos- ing discarded smartphones to minimize carbon

    Jennifer Switzer, Gabriel Marcano, Ryan Kastner, and Pat Pannuto. Junkyard computing: Repurpos- ing discarded smartphones to minimize carbon. InProceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume...

    2023

  6. [14]

    Jaylen Wang, Daniel S. Berger, Fiodar Kazhamiaka, Celine Irvene, Chaojie Zhang, Esha Choukse, Kali Frost, Rodrigo Fonseca, Brijesh Warrier, Chetan Bansal, Jonathan Stern, Ricardo Bianchini, and Akshitha Sriraman. Designing cloud servers for lower carbon. In2024 ACM/IEEE 51st A...

    2024

  7. [15]

    Carbon-aware computing for datacenters.IEEE Transac- tions on Power Systems, 38(2):1270–1280, 2023

    Ana Radovanovi´c, Ross Koningstein, Ian Schneider, Bokan Chen, Alexandre Duarte, Binz Roy, Diyue Xiao, Maya Haridasan, Patrick Hung, Nick Care, Saurav Talukdar, Eric Mullen, Kendal Smith, MariEllen Cottman, and Walfredo Cirne. Carbon-aware computing for datacenters.IEEE Transa...

    2023

  8. [16]

    Reducing the carbon impact of generative AI inference (today and in 2035)

    Andrew A Chien, Liuzixuan Lin, Hai Nguyen, Varsha Rao, Tristan Sharma, and Rajini Wijayawar- dana. Reducing the carbon impact of generative AI inference (today and in 2035). InProceedings of the 2nd Workshop on Sustainable Computer Systems, HotCarbon ’23, New York, NY, USA, 20...

    2023

  9. [17]

    Hanafy, Qianlin Liang, Noman Bashir, Abel Souza, David Irwin, and Prashant Shenoy

    Walid A. Hanafy, Qianlin Liang, Noman Bashir, Abel Souza, David Irwin, and Prashant Shenoy. Going green for less green: Optimizing the cost of reducing cloud carbon emissions. InProceedings of the 29th ACM International Conference on Architectural Support for Programming Langu...

    2024

  10. [18]

    2024 H1 semi-annual monitoring report (Intel Ocotillo facility).https://www.exploreintel

    Intel. 2024 H1 semi-annual monitoring report (Intel Ocotillo facility).https://www.exploreintel. com/ocotillo, 2024

    2024

  11. [19]

    Interagency monitoring of protected visual environments

    UC Davis Air Quality Research Center. Interagency monitoring of protected visual environments. https://airquality.ucdavis.edu/improve

  12. [20]

    U.S. EPA. Research on health effects from air pollution.https://www.epa.gov/air-research/ research-health-effects-air-pollution

  13. [21]

    Long-term exposure to PM 2.5 and cognitive decline: A longitudinal population-based study.Journal of Alzheimer’s Disease, 80(2):591–599, 2021

    Giulia Grande, Jing Wu, Petter LS Ljungman, Massimo Stafoggia, Tom Bellander, and Debora Rizzuto. Long-term exposure to PM 2.5 and cognitive decline: A longitudinal population-based study.Journal of Alzheimer’s Disease, 80(2):591–599, 2021

    2021

  14. [22]

    U.S. EPA. User’s manual for the co-benefits risk assessment (COBRA) screening model.https: //www.epa.gov/cobra/users-manual-co-benefits-risk-assessment-cobra-screening-model. 20

  15. [23]

    Estimates of global mortality burden associated with short-term exposure to fine particulate matter (PM2.5).The Lancet Planetary Health, 8(3):e146–e155, 2024

    Wenhua Yu, Rongbin Xu, Tingting Ye, Michael J Abramson, Lidia Morawska, Bin Jalaludin, Fay H Johnston, Sarah B Henderson, Luke D Knibbs, Geoffrey G Morgan, et al. Estimates of global mortality burden associated with short-term exposure to fine particulate matter (PM2.5).The La...

    2024

  16. [24]

    U.S. EPA. What is cross-state air pollution?https://www.epa.gov/Cross-State-Air-Pollution/ what-cross-state-air-pollution

  17. [25]

    Trivikrama Rao

    Jian Zhang and S. Trivikrama Rao. The role of vertical mixing in the temporal evolution of ground- level ozone concentrations.Journal of Applied Meteorology, 38(12):1674–1691, 1999

    1999

  18. [26]

    Guidance for evaluating human health impacts in environmental assessment: Air quality.https://iaac-aeic.gc.ca/050/documents/p80054/119376E.pdf, 2016

    Health Canada. Guidance for evaluating human health impacts in environmental assessment: Air quality.https://iaac-aeic.gc.ca/050/documents/p80054/119376E.pdf, 2016

    2016

  19. [27]

    World Health Organization. Air pollution is responsible for 6.7 million premature deaths every year.https://www.who.int/teams/environment-climate-change-and-health/ air-quality-and-health/health-impacts/types-of-pollutants

  20. [28]

    Ambient (outdoor) air pollution.https://www.who.int/news-room/ fact-sheets/detail/ambient-(outdoor)-air-quality-and-health, 2024

    World Health Organization. Ambient (outdoor) air pollution.https://www.who.int/news-room/ fact-sheets/detail/ambient-(outdoor)-air-quality-and-health, 2024

    2024

  21. [29]

    Global burden of disease 2021: Find- ings from the GBD 2021 study.https://www.healthdata.org/research-analysis/library/ global-burden-disease-2021-findings-gbd-2021-study, May 2024

    Institute for Health Metrics and Evaluation (IHME). Global burden of disease 2021: Find- ings from the GBD 2021 study.https://www.healthdata.org/research-analysis/library/ global-burden-disease-2021-findings-gbd-2021-study, May 2024

    2021

  22. [30]

    U.S. EPA. Human health and environmental impacts of the electric power sector.https://www.epa. gov/power-sector/human-health-environmental-impacts-electric-power-sector

  23. [31]

    U.S. EPA. Power plants and neighboring communities.https://www.epa.gov/power-sector/ power-plants-and-neighboring-communities

  24. [32]

    Mortality risk from united states coal electricity generation.Science, 382(6673):941–946, 2023

    Lucas Henneman, Christine Choirat, Irene Dedoussi, Francesca Dominici, Jessica Roberts, and Cor- win Zigler. Mortality risk from united states coal electricity generation.Science, 382(6673):941–946, 2023

    2023

  25. [33]

    The costs to health and the environment from indus- trial air pollution in Europe — 2024 update.https://www.eea.europa.eu/publications/ the-cost-to-health-and-the, 2025

    European Environment Agency. The costs to health and the environment from indus- trial air pollution in Europe — 2024 update.https://www.eea.europa.eu/publications/ the-cost-to-health-and-the, 2025

    2024

  26. [34]

    Diesel pollution from data centers.https://ecology.wa.gov/ air-climate/air-quality/data-centers

    Washington Department of Ecology. Diesel pollution from data centers.https://ecology.wa.gov/ air-climate/air-quality/data-centers

  27. [35]

    White House

    The U.S. White House. Executive order on advancing United States leadership in artificial intelli- gence infrastructure.https://www.federalregister.gov/documents/2025/01/17/2025-01395/ advancing-united-states-leadership-in-artificial-intelligence-infrastructure, Jan- uary 2025

    2025

  28. [36]

    Model information.https://github.com/meta-llama/llama-models/blob/main/ models/llama3_1/MODEL_CARD.md

    Meta Llama. Model information.https://github.com/meta-llama/llama-models/blob/main/ models/llama3_1/MODEL_CARD.md

  29. [37]

    Sustainability report.https://sustainability.atmeta.com/ 2024-sustainability-report/, 2024

    Meta. Sustainability report.https://sustainability.atmeta.com/ 2024-sustainability-report/, 2024

    2024

  30. [38]

    Investing in the rising data center economy.White Paper, January 2023

    McKinsey. Investing in the rising data center economy.White Paper, January 2023

    2023

  31. [39]

    U.S. EPA. Co-benefits risk assessment health impacts screening and mapping tool (COBRA).https: //cobra.epa.gov/

  32. [40]

    National Center for Health Statistics

    U.S. National Center for Health Statistics. Factstats: Asthma.https://www.cdc.gov/nchs/fastats/ asthma.htm. 21

  33. [41]

    Vehicles registered by county.https://www.dmv.ca

    California DMV. Vehicles registered by county.https://www.dmv.ca. gov/portal/dmv-research-reports/research-development-data-dashboards/ vehicles-registered-by-county/

  34. [42]

    Report to the Governor and the General Assembly of Virginia: Data centers in Virginia (JLARC report 158), December 2024

    Virginia Joint Legislative Audit and Review Commission. Report to the Governor and the General Assembly of Virginia: Data centers in Virginia (JLARC report 158), December 2024

    2024

  35. [43]

    Health risks from diesel emissions in the Quincy area.Air Quality Program Report (Publication 20-02-019), August 2020

    Washington Department of Ecology. Health risks from diesel emissions in the Quincy area.Air Quality Program Report (Publication 20-02-019), August 2020

    2020

  36. [44]

    U.S. EPA. Use of backup generators to maintain the reliability of the electric grid.https://www.epa.gov/system/files/documents/2025-05/ rice-memo-on-duke-energy-regulatory-interpretation-04_17_25.pdf, May 2025

    2025

  37. [45]

    Donahue, Allen L

    Neil M. Donahue, Allen L. Robinson, and Spyros N. Pandis. Atmospheric organic particulate matter: From smoke to secondary organic aerosol.Atmospheric Environment, 43(1):94–106, 2009

    2009

  38. [46]

    U.S. EPA. Benefits and costs of the Clean Air Act 1990-2020, the sec- ond prospective study.https://www.epa.gov/clean-air-act-overview/ benefits-and-costs-clean-air-act-1990-2020-second-prospective-study, March 2011

    1990

  39. [47]

    U.S. EPA. Summary of the Clean Air Act.https://www.epa.gov/laws-regulations/ summary-clean-air-act

  40. [48]

    U.S. EPA. National ambient air quality standards (NAAQS) table.https://www.epa.gov/ criteria-air-pollutants/naaqs-table

  41. [49]

    Laws and regulations.https://ww2.arb.ca.gov/resources/ documents/laws-and-regulations

    California Air Resources Board. Laws and regulations.https://ww2.arb.ca.gov/resources/ documents/laws-and-regulations

  42. [50]

    State of the air.Report, 2024

    American Lung Association. State of the air.Report, 2024

    2024

  43. [51]

    Ambient air pollution attributable deaths.https://www.who.int/data/ gho/data/indicators/indicator-details/GHO/ambient-air-pollution-attributable-deaths, 2024

    World Health Organization. Ambient air pollution attributable deaths.https://www.who.int/data/ gho/data/indicators/indicator-details/GHO/ambient-air-pollution-attributable-deaths, 2024

    2024

  44. [52]

    WHO global air quality guidelines.https://www.who.int/ publications/i/item/9789240034228

    World Health Organization. WHO global air quality guidelines.https://www.who.int/ publications/i/item/9789240034228

  45. [53]

    U.S. EPA. Projection of counties with monitors that would not meet in 2032.https://www.epa. gov/system/files/documents/2024-02/2024-pm-naaqs-final-2032-projections-map.pdf, February 2024

    2024

  46. [54]

    Stevens and Supreme Court of The United States

    John P. Stevens and Supreme Court of The United States. U.S. reports: Massachusetts v. EPA, 549 U.S. 497.The Library of Congress, 2007

    2007

  47. [55]

    U.S. EPA. Climate change and human health.https://www.epa.gov/climateimpacts/ climate-change-and-human-health

  48. [56]

    Standards and guidance.https://ghgprotocol.org/

    Greenhouse Gas Protocol. Standards and guidance.https://ghgprotocol.org/

  49. [57]

    U.S. EPA. Controlling air pollution from stationary engines.https://www.epa.gov/ stationary-engines

  50. [58]

    Fact sheet on emergency backup generators.https://www.aqmd

    California Air Resources Board. Fact sheet on emergency backup generators.https://www.aqmd. gov/home/permits/emergency-generators

  51. [59]

    Explaining the Uptime Institute’s tier classification system (April 2021 update).https://journal.uptimeinstitute.com/ explaining-uptime-institutes-tier-classification-system/

    Uptime Institute. Explaining the Uptime Institute’s tier classification system (April 2021 update).https://journal.uptimeinstitute.com/ explaining-uptime-institutes-tier-classification-system/. 22

    2021

  52. [60]

    Issued air permits for data centers.https://www

    Virginia Department of Environmental Quality. Issued air permits for data centers.https://www. deq.virginia.gov/permits/air/issued-air-permits-for-data-centers

  53. [61]

    Hanna Pampaloni. Heat wave prompts increased data center generator use; turner pushes for tier 4 upgrades.https://www.loudounnow.com/news/ heat-wave-prompts-increased-data-center-generator-use-turner-pushes-for-tier-4-upgrades/ article_60a48bda-dc50-4d1a-8b16-399cd4340350.html...

    2025

  54. [62]

    Data centers, diesel generators and air quality – PEC web map.https://www.pecva.org/uncategorized/ data-centers-diesel-generators-and-air-quality-pec-web-map/

    Piedmont Environmental Council. Data centers, diesel generators and air quality – PEC web map.https://www.pecva.org/uncategorized/ data-centers-diesel-generators-and-air-quality-pec-web-map/

  55. [63]

    Government of Canada. Wet cooling towers: Guide to reporting.https://www.canada.ca/en/ environment-climate-change/services/national-pollutant-release-inventory/report/ sector-specific-tools-calculate-emissions/wet-cooling-tower-particulate-guide.html

  56. [64]

    Wexler, Chris D

    Anthony S. Wexler, Chris D. Wallis, Patrick Chuang, and Mason Leandro. Assessing particulate emis- sions from power plant cooling towers.California Energy Commission Final Project Report (CEC-500- 2023-048), July 2013

    2023

  57. [65]

    Reuters. Data center reliance on fossil fuels may delay clean-energy tran- sition.https://www.reuters.com/technology/artificial-intelligence/ how-ai-cloud-computing-may-delay-transition-clean-energy-2024-11-21/, November 2024

    2024

  58. [66]

    Integrated resource plan.https://www.dominionenergy

    Virginia Electric and Power Company. Integrated resource plan.https://www.dominionenergy. com/-/media/pdfs/global/company/IRP/2024-IRP-w_o-Appendices.pdf, October 2024

    2024

  59. [67]

    Energy Information Administration

    U.S. Energy Information Administration. Annual energy outlook 2025.https://www.eia.gov/ outlooks/aeo

    2025

  60. [68]

    Electricity mix.Our World in Data, 2024

    Hannah Ritchie and Pablo Rosado. Electricity mix.Our World in Data, 2024

    2024

  61. [69]

    New nuclear clean energy agreement with Kairos Power.https://blog.google/ outreach-initiatives/sustainability/google-kairos-power-nuclear-energy-agreement/, October 2024

    Google. New nuclear clean energy agreement with Kairos Power.https://blog.google/ outreach-initiatives/sustainability/google-kairos-power-nuclear-energy-agreement/, October 2024

    2024

  62. [70]

    To land Meta’s massive$10 billion data center, Louisiana pulled out all the stops

    CNBC. To land Meta’s massive$10 billion data center, Louisiana pulled out all the stops. will it be worth it?https://www.cnbc.com/2025/06/25/ meta-massive-data-center-louisiana-cost-jobs-energy-use.html, June 2025

    2025

  63. [71]

    The Associated Press. OpenAI shows off Stargate AI data center in Texas and plans 5 more elsewhere with Oracle, Softbank.https://apnews.com/article/ openai-stargate-oracle-data-center-0b3f4fa6e8d8141b4c143e3e7f41aba1, September 2025

    2025

  64. [72]

    DynamoLLM: De- signing LLM inference clusters for performance and energy efficiency

    Jovan Stojkovic, Chaojie Zhang, Inigo Goiri, Josep Torrellas, and Esha Choukse. DynamoLLM: De- signing LLM inference clusters for performance and energy efficiency. InIEEE International Symposium on High-Performance Computer Architecture (HPCA), 2025

    2025

  65. [73]

    E-waste challenges of generative artificial intelligence.Nature Computational Science, 4:818–823, October 2024

    Peng Wang, Ling-Yu Zhang, Asaf Tzachor, and Wei-Qiang Chen. E-waste challenges of generative artificial intelligence.Nature Computational Science, 4:818–823, October 2024

    2024

  66. [74]

    U.S. EPA. Semiconductor industry.https://www.epa.gov/eps-partnership/ semiconductor-industry

  67. [75]

    Generative AI: The next S-curve for the semiconductor industry?White Paper, March 2024

    McKinsey. Generative AI: The next S-curve for the semiconductor industry?White Paper, March 2024

    2024

  68. [76]

    U.S. EPA. Avoided emissions and generation tool (A VERT).https://www.epa.gov/avert

  69. [77]

    WattTime.https://watttime.org/. 23

  70. [78]

    Environmental sustainability report.https://www.microsoft.com/en-us/ corporate-responsibility/sustainability/report, 2024

    Microsoft. Environmental sustainability report.https://www.microsoft.com/en-us/ corporate-responsibility/sustainability/report, 2024

    2024

  71. [79]

    Life-cycle emissions of AI hardware: A cradle-to-grave approach and generational trends, 2025

    Ian Schneider, Hui Xu, Stephan Benecke, David Patterson, Keguo Huang, Parthasarathy Ran- ganathan, and Cooper Elsworth. Life-cycle emissions of AI hardware: A cradle-to-grave approach and generational trends, 2025

    2025

  72. [80]

    U.S. EPA. Air quality dispersion modeling.https://www.epa.gov/scram/ air-quality-dispersion-modeling

  73. [81]

    McNider and Arastoo Pour-Biazar

    Richard T. McNider and Arastoo Pour-Biazar. Meteorological modeling relevant to mesoscale and regional air quality applications: A review.Journal of the Air & Waste Management Association, 70(1):2– 43, 2020

    2020

  74. [82]

    Tessum, Jason D

    Christopher W. Tessum, Jason D. Hill, and Julian D. Marshall. InMAP: A model for air pollution interventions.PloS ONE, 12(4):e0176131, 2017

    2017

  75. [83]

    Baker, Heather Simon, Barron Henderson, Colby Tucker, David Cooley, and Emma Zinsmeis- ter

    Kirk R. Baker, Heather Simon, Barron Henderson, Colby Tucker, David Cooley, and Emma Zinsmeis- ter. Source–receptor relationships between precursor emissions and O3 and PM2.5 air pollution im- pacts.Environmental Science & Technology, 57(39):14626–14637, 2023

    2023

  76. [84]

    Schwartz

    Qian Di, Yan Wang, Antonella Zanobetti, Yun Wang, Petros Koutrakis, Christine Choirat, Francesca Dominici, and Joel D. Schwartz. Air pollution and mortality in the medicare population.New England Journal of Medicine, 376(26):2513–2522, 2017

    2017

  77. [85]

    U.S. EPA. Publications that cite COBRA.https://www.epa.gov/cobra/publications-cite-cobra

  78. [86]

    Jean Schmitt, Marianne Hatzopoulou, Amir FN Abdul-Manan, Heather L MacLean, and I Daniel Posen. Health benefits of US light-duty vehicle electrification: Roles of fleet dynamics, clean electricity, and policy timing.Proceedings of the National Academy of Sciences, 121(43):e232...

    2024

  79. [87]

    The environmental burden of the United States’ bitcoin mining boom.https://pubmed.ncbi.nlm.nih.gov/39502776/, 2024

    Gianluca Guidi, Francesca Dominici, Nat Steinsultz, Gabriel Dance, Lucas Henneman, Henry Richardson, Edgar Castro, Falco J Bargagli-Stoffi, and Scott Delaney. The environmental burden of the United States’ bitcoin mining boom.https://pubmed.ncbi.nlm.nih.gov/39502776/, 2024

    2024

  80. [88]

    Hennessy, Jacques A

    Eleanor M. Hennessy, Jacques A. de Chalendar, Sally M. Benson, and Inˆes ML Azevedo. Distributional health impacts of electricity imports in the United States.Environmental Research Letters, 17(6):064011, 2022

    2022

  81. [89]

    National Park Service

    U.S. National Park Service. Where does air pollution come from?https://www.nps.gov/subjects/ air/sources.htm

  82. [90]

    U.S. EPA. Clean Air Act vehicle and engine enforcement case resolutions.https://www.epa.gov/ enforcement/clean-air-act-vehicle-and-engine-enforcement-case-resolutions

  83. [91]

    U.S. EPA. Final rule: Multi-pollutant emissions standards for model years 2027 and later light-duty and medium-duty vehicles. https://www.epa.gov/regulations-emissions-vehicles-and-engines/ final-rule-multi-pollutant-emissions-standards-model, April 2024

    2027

  84. [92]

    Introducing Llama 3.1: Our most capable models to date.https://ai.meta.com/blog/ meta-llama-3-1/

    Meta. Introducing Llama 3.1: Our most capable models to date.https://ai.meta.com/blog/ meta-llama-3-1/

  85. [93]

    Deepseek-v3 technical report.arXiv 2412.19437, 2024

    DeepSeek-AI. Deepseek-v3 technical report.arXiv 2412.19437, 2024

    2024 arXiv

  86. [94]

    Once for all: Train one network and specialize it for efficient deployment

    Han Cai, Chuang Gan, and Song Han. Once for all: Train one network and specialize it for efficient deployment. InICLR, 2019. 24

    2019

  87. [95]

    Cutting the electric bill for internet-scale systems

    Asfandyar Qureshi, Rick Weber, Hari Balakrishnan, John Guttag, and Bruce Maggs. Cutting the electric bill for internet-scale systems. InProceedings of the ACM SIGCOMM 2009 Conference on Data Communication, SIGCOMM ’09, page 123–134, New York, NY, USA, 2009. Association for Com...

    2009

  88. [96]

    Islam, Kishwar Ahmed, Hong Xu, Nguyen H

    Mohammad A. Islam, Kishwar Ahmed, Hong Xu, Nguyen H. Tran, Gang Quan, and Shaolei Ren. Ex- ploiting spatio-temporal diversity for water saving in geo-distributed data centers.IEEE Transactions on Cloud Computing, 6(3):734–746, 2018

    2018

  89. [97]

    Towards environmentally equitable AI via geographical load balancing

    Pengfei Li, Jianyi Yang, Adam Wierman, and Shaolei Ren. Towards environmentally equitable AI via geographical load balancing. Ine-Energy, 2024

    2024

  90. [98]

    Signal: Health damage.https://watttime.org/data-science/data-signals/ health-damage/

    WattTime. Signal: Health damage.https://watttime.org/data-science/data-signals/ health-damage/

  91. [99]

    ElectricityEmissions.jl: A framework for the comparison of carbon intensity signals.arXiv 2411.06560, 2024

    Joe Gorka, Noah Rhodes, and Line Roald. ElectricityEmissions.jl: A framework for the comparison of carbon intensity signals.arXiv 2411.06560, 2024

    2024 arXiv

  92. [100]

    U.S. EPA. Estimating the health benefits per kilowatt-hour of energy ef- ficiency and renewable energy.https://www.epa.gov/statelocalenergy/ estimating-health-benefits-kilowatt-hour-energy-efficiency-and-renewable-energy, 2024

    2024

  93. [101]

    Institute for Energy Research. EPA proposes exorbitant estimate for the so- cial cost of carbon.https://www.instituteforenergyresearch.org/regulation/ epa-proposes-exorbitant-estimate-for-the-social-cost-of-carbon/, 2022

    2022

  94. [102]

    U.S. EIA. Electric power plants, capacity, generation, fuel consumption, sales, prices and customers. https://www.eia.gov/electricity/data.php, 2023

    2023

  95. [103]

    U.S. EPA. About the U.S. electricity system and its impact on the environment.https://www.epa. gov/energy/about-us-electricity-system-and-its-impact-environment

  96. [104]

    Department of Transportation

    U.S. Department of Transportation. Estimated U.S. average vehicle emissions rates per vehicle by vehicle type using gasoline and diesel.National Transportation Statistics Table 4-43, June 2024

    2024

  97. [105]

    Energy Information Administration

    U.S. Energy Information Administration. Annual energy outlook 2023.https://www.eia.gov/ outlooks/aeo

    2023

  98. [106]

    U.S. EPA. Arizona nonattainment/maintenance status for each county by year for all criteria pollu- tants.https://www3.epa.gov/airquality/greenbook/anayo_az.html, 2024

    2024

  99. [107]

    Ocotillo campus.https://www.exploreintel.com/ocotillo, 2024

    Intel. Ocotillo campus.https://www.exploreintel.com/ocotillo, 2024

    2024

  100. [108]

    2023-24 corporate responsibility report, 2024

    Intel. 2023-24 corporate responsibility report, 2024

    2023

  101. [109]

    counties

    U.S. Energy Information Administration (EIA). EIA open data.https://www.eia.gov/opendata/. 25 Appendix A Modeling Details We describe the evaluation methodology used for our empirical analysis. We use the latest COBRA (Desktop v5.1, as of October 2024) provided by the U.S. EPA...

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.