Original Paper:

Guidi, G., Dominici, F., Steinsultz, N., Dance, G., Henneman, L., Richardson, H., ... & Delaney, S. (2025). The environmental burden of the United States’ bitcoin mining boom. Nature Communications, 16(1), 2970. https://www.nature.com/articles/s41467-025-58287-3

While 'The environmental burden of the United States’ bitcoin mining boom' aims to provide a critical analysis of Bitcoin mining's impact on PM2.5 pollution, our critique highlights substantial shortcomings that undermine its validity and usefulness.

These include selective use of data, a flawed method of attributing emissions, the inappropriate application of marginal emissions calculations, and an over-reliance on news media and a small number of selectively chosen peer-reviewed studies that support a specific narrative.

These shortcomings result in the article exaggerating the impact of Bitcoin mining on PM2.5 pollution.  This is unfortunate, given the growing body of other, more balanced, research on the environmental costs and benefits of bitcoin mining.

1. Deficient Academic Source Base

The literature that forms the scholarly basis for this study contains serious omissions and is not of the standard expected of a journal of the quality of Nature Communications.  

Specific shortcomings include:

1.    Overuse of Non-Scholarly Sources:

●      News articles (NYT, Forbes, EcoWatch, Bloomberg) dominate, providing anecdotal context but lacking peer-reviewed validation.

●      Government/policy reports (EPA, EIA) offer regulatory framing but do not engage with Bitcoin-specific energy debates.

2.    Lack of Engagement with Foundational Bitcoin Scholarship

●      Only 3 peer-reviewed studies directly address Bitcoin mining (Jiang et al., 2021; Jones et al., 2022; Henneman et al., 2023).

●      No engagement with foundational literature on Bitcoin’s potential to accelerate renewable energy adoption or stabilize grids (e.g., Lal et al., 2023; Bruno et al., 2023, Menati et al., 2023, Rudd et al., 2024, Cambridge’s mining reports).

3.     Major Energy Trackers Ignored

●      Major Bitcoin mining energy trackers estimate that Bitcoin mining uses primarily electricity from low emission sources and observe that mining operations often act as flexible load resources, absorbing excess renewable generation.  The Digital Assets Research Institute (2024), for example, has estimated that 56.7% of Bitcoin mining uses sustainable energy.  

4.    Methodological Gaps:

●      No engagement with critiques of marginal emissions models (e.g., WattTime’s marginal operating emissions rate [MOER] limitations) or alternative accounting methods (e.g., GHG Protocol Scope 2 Guidance, contractual emissions etc).

In summary, the references cited lack the depth of peer-reviewed Bitcoin-energy research expected in academic work. Reliance on news and government sources weakens scholarly rigor, as critical counterarguments and methodological debates are omitted.

Impact: These omissions create a one-sided narrative, framing Bitcoin mining solely as a fossil fuel-driven industry rather than a potential enabler of decarbonization (a thesis supported by no fewer than 15 studies on the subject of Bitcoin and energy).

Given Nature Communications' reputation for publishing rigorously peer-reviewed research, it is notable that the reference list leans heavily on non-scholarly sources and lacks engagement with foundational Bitcoin-energy literature. This may reflect the interdisciplinary or emerging nature of the topic, though such an approach could be perceived as inconsistent with the journal’s typical standards for contextualizing findings within prior academic research.

2. Flawed Emissions Attribution

The study attributes emissions to Bitcoin mining based on balancing authority regions, not direct contractual agreements. This assumes fossil plants are the marginal suppliers, which is often inaccurate:

• Proximity Fallacy: Emissions are linked to distant fossil plants (e.g., a Kentucky coal plant powering a North Carolina mine), but grids increasingly prioritize renewables for marginal demand.
• Renewable Sourcing: Many mines explicitly purchase renewables (e.g., Marathon in Hansford County (Marathon Digital Holdings, 2024)), but the methodology ignores such contracts.

Consequence: Emissions are misallocated to Bitcoin mining operations even when their energy is green.

3. Misuse of Marginal Emissions Calculations

The study employs WattTime’s marginal operating emissions rate (MOER) to attribute emissions to Bitcoin mines. However, this has a number of serious drawbacks:

• Short-Term vs. Long-Term: MOER is designed to estimate short-term grid responses (e.g., hourly demand shifts), not long-term structural changes. Applying it to annual data risks overestimating emissions, as grids adapt over time (e.g., adding renewables or storage).
• Departure from Accepted Emissions Accounting: The MOER method of attributing emissions is inconsistent with international norms for attributing emissions, such as the GHG Protocol Scope 2 Guidance, and tends to allocate far more emissions to the newest energy user.  
• Bitcoin Mining is Atypical: Bitcoin mining is an atypical user of electricity as miners will commonly reduce load during periods of high demand due to incentives provided by demand response programs. It is, therefore, one of the least suitable load types to model using MOER.
• Proprietary Model: WattTime’s methodology is opaque, making replication impossible. The lack of transparency undermines scientific rigor.

Example: If a Bitcoin mine contracts with a wind farm, MOER might still attribute emissions to fossil plants due to grid-balancing dynamics, even if the mine’s direct consumption is renewable.

4. Selective use of Data

The study focuses on 34 of 137 U.S. Bitcoin mining operations, excluding:

• Renewable-Collocated Mines: The study excludes many US-based Bitcoin mining operations co-located with renewables, including large ones such as MARA’s 114 MW windfarm, or Crusoe’s 200 MW flare-gas mitigation Bitcoin mining datacenters, skewing results toward on-grid operations, which tend to have a relatively higher fossil-fuel mix.
• Geographic Bias: Texas forms less than one-third of the sample (10/34 mining operations), yet Texas’ grid (ERCOT) has 50% of US hashrate, and a higher penetration of renewable energy.
• Result: The analysis inflates Bitcoin’s fossil dependency by omitting mining operations using renewable energy.

5. Biased Language

The article uses emotive terms (e.g., “toxic PM2.5,” “prodigious electricity use”) without contextualizing Bitcoin’s energy use against other industries. For example:

• Comparisons: Global data centers consume ~460 TWh/year (2022) vs. Bitcoin’s ~150 TWh/year, yet the article singles out Bitcoin for criticism.
• Direct vs. Indirect Emissions: The methodology could apply to any electrified industry (e.g., electric vehicles, cloud computing), but Bitcoin is uniquely framed as a polluter.

If the author is genuinely concerned about reducing PM2.5 pollution it would be logical to call for more efforts to regulate the thermal power stations that produce PM2.5, rather than focusing primarily on Bitcoin mining, which is just one of many electricity users.

This is especially problematic given that flexible load data centers such as Bitcoin mining have been shown to help grid operators place more renewable energy on the grid. In contrast, inflexible data centers, such as those that process inference AI workloads, require more fossil-fuel-based generation (Rhodes et al., 2021).

Implication: The language used in the article risks sensationalizing Bitcoin’s impact while ignoring systemic grid challenges.

6. Methodological Limitations

• Dated Emissions Data: The study uses EPA’s 2018 eGRID data but analyzes 2022–2023 operations. Power plant retirements and renewable additions (e.g., U.S. solar capacity grew 50% in 2023) are ignored.
• InMAP Model Uncertainty: The model’s coarse rural resolution and reliance on 2005 meteorological data reduce accuracy, especially for dispersed pollution.

7. Independent Observations

• Policy Context: The study advocates for federal regulation but overlooks existing efforts. For example, Texas’ demand-response programs reward Bitcoin miners for shutting down during grid stress, reducing fossil reliance.
• Health Risk Overstatement: PM2.5 increases of 0.1–0.5 µg/m³ are statistically significant but small compared to the national average of ~10 µg/m³. The article does not contextualize this incremental risk.

Conclusion

While 'The environmental burden of the United States’ bitcoin mining boom' attempts to provide a critical analysis of Bitcoin mining's impact on PM2.5 pollution, its narrow scope, methodological flaws, and skewed list of citations undermine its validity and usefulness.

As detailed in our critique, these shortcomings prevent the article from contributing meaningfully to the growing body of balanced research on the environmental costs and benefits of Bitcoin. To rectify these issues, future research should:

1. Include a full dataset encompassing renewable-colocated mining operations.
2. Use the most up-to-date data.
3. Employ granular, contract-specific emissions accounting.
4. Recognize and contextualize the existing literature on Bitcoin’s potential for grid stabilization and renewable energy integration.
5. Be cognizant of the limitations of different emissions accounting models and be sure to use the appropriate model for the application.

By incorporating these critical improvements, future research can foster a more scientifically sound debate, ultimately guiding informed and balanced policy-making that considers both the environmental impacts and potential benefits of Bitcoin mining.

References

Bruno, A., Weber, P., & Yates, A. J. (2023). Can Bitcoin mining increase renewable electricity capacity? Resource and Energy Economics, 74, 101376. https://doi.org/10.1016/j.reseneeco.2023.101376

Collins, J., & Dewhurst, R. D. (2024). Runaway citations and the persistence of Bitcoin misinformation. Digital Assets Research Institute.​ https://www.da-ri.org/articles/working-paper-runaway-citations-and-the-persistence-of-bitcoin-misinformation

Digital Assets Research Institute (2024). BEEST Model. Retrieved from https://woocharts.com/esg-bitcoin-mining-sustainability/​

‍Guidi, G., Dominici, F., Steinsultz, N. et al. The environmental burden of the United States’ bitcoin mining boom. Nat Commun 16, 2970 (2025). https://doi.org/10.1038/s41467-025-58287-3

Lal, A., Zhu, J., & You, F. (2023). From mining to mitigation: How Bitcoin can support renewable energy development and climate action. ACS Sustainable Chemistry & Engineering, 11(45), 16330–16340. ​https://doi.org/10.1021/acssuschemeng.3c05445

Marathon Digital Holdings. (2024, April 4). Marathon Holdings announces closing of Texas wind farm acquisition [Press release]. Marathon Digital Holdings. https://ir.mara.com/news-events/press-releases/detail/1389/mara-holdings-announces-closing-of-texas-wind-farm-acquisition

Menati, A., Zheng, X., Lee, K., Shi, R., Du, P., Singh, C., & Xie, L. (2023). High resolution modeling and analysis of cryptocurrency mining’s impact on power grids: Carbon footprint, reliability, and electricity price. Advances in Applied Energy, 10, 100136. ISSN 2666-7924. https://doi.org/10.1016/j.adapen.2023.100136

Rhodes, J. D., Deetjen, T., & Smith, C. (2021). Impacts of Large, Flexible Data Center Operations on the Future of ERCOT [White paper]. Lancium. Retrieved from https://lancium.com/wp-content/uploads/2022/06/Lancium_Flexible_Data_Center_Whitepaper_4.2022.pdf

Rudd, M. A., Jones, M., Sechrest, D., Batten, D., & Porter, D. (2024). An integrated landfill gas-to-energy and Bitcoin mining framework. Journal of Cleaner Production, 472, 143516.​ https://doi.org/10.1016/j.jclepro.2024.143516