Rethinking the Environmental Costs of Training AI — Why We Should Look Beyond Hardware

1.

Ever for the reason that Forties, when the primary digital computer systems had been invented, scientists have at all times dreamed of making machines as good as people, what now grew to become Synthetic Intelligence (AI). Quick ahead to November 2022, when ChatGPT — an AI mannequin able to listening and answering immediately — was launched, it felt like a dream come true. Afterward, tons of of recent AI fashions have rushed into the race (check out the timeline right here). In the present day, each single day, one billion messages are despatched via ChatGPT (OpenAI Newsroom, 2024), highlighting the speedy AI adoption by customers. But, few folks cease to ask: What are the environmental prices behind this new comfort?

Earlier than customers can ask AI questions, these fashions should first be skilled. Coaching is the method the place fashions, or algorithms, are fed datasets and attempt to discover one of the best match. Think about a easy regression y = ax + b: coaching means feeding the algorithm x and y values and permitting it to search out one of the best parameters a and b. In fact, AI fashions sometimes wouldn’t be so simple as a linear regression. They’d include tons of parameters, thus requiring huge quantities of computation and datasets. Furthermore, they would wish to run a considerable quantity of specialised {hardware} that may deal with that sheer quantity of computation and complexity. All of that mixed made AI eat far more power than conventional software program.

As well as, AI coaching requires a secure and uninterrupted power provide, which primarily comes from non-renewable power sources like pure fuel or coal-based, as a result of photo voltaic and wind power can fluctuate based mostly on climate circumstances (Calvert, 2024). Furthermore, as a result of excessive depth of power use, knowledge facilities — buildings that retailer AI fashions — warmth up quickly, emitting vital carbon footprints and requiring massive quantities of water for cooling. Due to this fact, AI fashions have broad environmental impacts that embrace not solely power utilization but in addition water consumption and carbon emissions.

Sadly, there’s not a lot official and disclosed knowledge relating to power, water, and carbon footprints of AI fashions. The general public stays largely unaware of those environmental impacts and thus has not created robust stress or motivations for tech corporations to take extra systematic adjustments. Moreover, whereas some enhancements have been made — particularly in {hardware} power effectivity — there stays little systematic or coordinated effort to successfully scale back the general environmental impacts of AI. Due to this fact, I hope to enhance public consciousness of those hidden environmental prices and to discover whether or not latest enhancements in power effectivity are substantial. Extra significantly, I’m searching for to deal with two analysis questions on this research:

RQ1: Is there a major relationship between AI fashions’ architectural and {hardware} decisions and their useful resource consumption throughout coaching?

RQ2: Has AI coaching change into energy-efficient over time?

2. Strategies: 

The paper used a dataset known as Notable AI Fashions from Epoch AI (Epoch AI, 2025), a analysis institute that investigates the developments of AI improvement. The fashions included had been both traditionally related or symbolize cutting-edge advances in AI. Every mannequin was recorded with key coaching data such because the variety of parameters, dataset dimension, complete compute, {hardware} sort, and {hardware} amount, all collected from varied sources, together with literature opinions, publications, and analysis papers. The dataset additionally reported the boldness degree for these attributes. To supply a dependable evaluation, I evaluated solely fashions with a confidence score of “Assured” or “Seemingly”.

As famous earlier, there was restricted knowledge relating to direct useful resource consumption. Luckily, the dataset authors have estimated Whole Energy Draw (in watts, or W) based mostly on a number of components, together with {hardware} sort, {hardware} amount, and another knowledge heart effectivity charges and overhead. You will need to word that energy and power are totally different: energy (W) refers back to the quantity of electrical energy used per unit of time, whereas power (in kilowatt-hours, or kWh) measures the overall cumulative electrical energy consumed over time.

Since this research investigated useful resource consumption and power effectivity through the coaching section of AI fashions, I constructed and estimated 4 environmental metrics: complete power used (kWh), complete water used (liters, or L), complete carbon emissions (kilograms of CO2e, or kgCO2e), and power effectivity (FLOPS/W, to be defined later).

a. Estimation strategies

First, this research estimated power consumption by deciding on fashions with out there complete energy draw (W) and coaching instances (hours). Power was computed as follows:

[text{Energy (kWh)} = frac{text{Total Power Draw (W)}}{1000} times text{Training Time (h)}]

Subsequent, water consumption and carbon emissions had been estimated by rearranging the formulation of two normal charges utilized in knowledge facilities: Water Utilization Effectiveness (WUE, in L/kWh) and Carbon Depth (CI, in kgCO2e/kWh):

[text{WUE (L/kWh)} = frac{text{Water (L)}}{text{Energy (kWh)}} Longrightarrow text{Water (L)} = text{WUE (L/kWh)} times text{Energy (kWh)}]

This research used the common WUE of 0.36 L/kWh in 2023, reported by Lawrence Berkeley Nationwide Laboratory (2024).

[mathrm{CI left( frac{mathrm{kgCO_2e}}{mathrm{kWh}} right)} = frac{mathrm{Carbon (kgCO_2e)}}{mathrm{Energy (kWh)}} Longrightarrow mathrm{Carbon (kgCO_2e)} = mathrm{CI left( frac{mathrm{kgCO_2e}}{mathrm{kWh}} right)} times mathrm{Energy (kWh)}]

This research used a median carbon depth of 0.548 kg CO₂e/kWh, reported by latest environmental analysis (Guidi et al, 2024).

Lastly, this research estimated power effectivity utilizing the FLOPS/W metric. A floating-point operation (FLOP) is a fundamental arithmetic operation (e.g., addition or multiplication) with decimal numbers. FLOP per second (FLOPS) measures what number of such operations a system can carry out every second, and is usually used to judge computing efficiency. FLOPS per Watt (FLOPS/W) measures how a lot computing efficiency is achieved per unit of energy consumed:

[text{Energy Efficiency (FLOPS/W)} = frac{text{Total Compute (FLOP)}}{text{Training Time (h)} times 3600 times text{Total Power Draw(W)}}]

You will need to word that FLOPS/W is usually used to measure hardware-level power effectivity. Nonetheless, it’s potential that the precise effectivity throughout AI coaching could also be totally different from the thereotical effectivity reported for the {hardware} used. I wish to examine whether or not any of the training-related components, past {hardware} alone, might contribute considerably to general power effectivity.

b. Evaluation strategies:

RQ1: Architectural and {Hardware} Decisions vs Useful resource Consumption

 Amongst power, water, and carbon consumption, I targeted on modeling power consumption, as each water and carbon are derived immediately from power utilizing mounted conversion charges and all three response variables shared equivalent distributions. In consequence, I consider we may safely assume that the best-fitting mannequin of power consumption might be utilized to water and carbon. Whereas the statistical fashions had been the identical, I might nonetheless report the outcomes of all three to quantify what number of kilowatt-hours of power, liters of water, and kilograms of carbon are wasted for each unit enhance in every vital issue. That approach, I hope to speak the environmental impacts of AI in a extra holistic, concrete, and tangible phrases.

Based mostly on Determine 1, the histogram of power confirmed excessive proper skew and the presence of some outliers. Due to this fact, I carried out a log transformation on power knowledge, aiming to stabilize variance and transfer the distribution nearer to normality (Fig. 2). A Shapiro-Wilk check confirmed the log-transformed power knowledge is roughly regular (p-value = 0.5). Based mostly on this, two kinds of distributions had been thought of: the Gaussian (regular) and the Gamma distribution. Whereas the Gaussian distribution is approriate for symmetric and regular knowledge, the Gamma distribution is extra fitted to optimistic, skewed knowledge — generally utilized in engineering modeling the place small values happen extra continuously than bigger values. For every distribution, the paper in contrast two approaches for incorporating the log transformation: immediately log reworking the response variable versus utilizing a log hyperlink perform inside a generalized linear mannequin (GLM). I recognized one of the best mixture of distribution and log method by evaluating their Akaike Info Criterion (AIC), diagnostic plots, together with prediction accuracy.

The candidate predictors included Parameters, Coaching Compute, Dataset Measurement, Coaching Time, {Hardware} Amount, and {Hardware} Sort. Structure-related variables comprised Parameters, Coaching Compute, and Dataset Measurement, whereas hardware-related variables consisted of {Hardware} Amount and {Hardware} Sort. Coaching Time didn’t fall neatly into both class however was included resulting from its central function in coaching AI fashions. After becoming all candidate predictors into the chosen GLM specification, I examined for multicollinearity to find out whether or not any variables needs to be excluded. Following this, I explored interplay phrases, as every useful resource consumption might not have responded linearly to every impartial variable. The next interactions had been thought of based mostly on area data and varied sources:

• Mannequin Measurement and {Hardware} Sort: Totally different {hardware} varieties have totally different reminiscence designs. The bigger and extra complicated the mannequin is, the extra reminiscence it requires (Bali, 2025). Power consumption might be totally different relying on how the {hardware} handles reminiscence calls for.
• Dataset Measurement and {Hardware} Sort: Equally, with totally different reminiscence designs, {hardware} might entry and skim knowledge at totally different knowledge dimension (Krashinsky et al, 2020). As dataset dimension will increase, power consumption can range relying on how the {hardware} handles massive volumes of information.
• Coaching Time with {Hardware} Amount: Working a number of {hardware} items on the identical time provides further overhead, like preserving every little thing in sync (HuggingFace, 2025). As coaching goes on, these coordination prices can develop and put extra pressure on the system, resulting in quicker power drain.
• Coaching Time with {Hardware} Sort: As coaching time will increase, power use might range throughout {hardware} varieties since some {hardware} varieties might handle warmth higher or preserve efficiency extra constantly over time, whereas others might decelerate or eat extra power.

RQ2: Power Effectivity over Time

The distribution of power effectivity was extremely skewed. Even after a log transformation, the distribution remained non-normal and overdispersed. To scale back distortion, I eliminated one excessive outlier with exceptionally excessive effectivity, because it was not a frontier mannequin and certain much less impactful. A Gamma GLM was then fitted utilizing Publication Date as the first predictor. If fashions utilizing the identical {hardware} exhibited huge variation in effectivity, it could recommend that different components past the {hardware} might contribute to those variations. Due to this fact, structure and {hardware} predictors from the primary analysis query can be used to evaluate which variables considerably affect power effectivity over time.

3. Outcomes

RQ1: Architectural and {Hardware} Decisions vs Useful resource Consumption

I finally used a Gamma GLM with a log hyperlink to mannequin useful resource consumption. This mix was chosen as a result of it had a decrease AIC worth (1780.85) than the Gaussian log-link mannequin (2005.83) and produced predictions that matched the uncooked knowledge extra carefully than fashions utilizing a log-transformed response variable. These log-transformed fashions generated predictions that considerably underestimated the precise knowledge on the unique scale (see this article on why log-transforming didn’t work in my case).

Structure Elements Don’t Maintain A lot Predictive Energy as {Hardware} Ones

After becoming all candidate explanatory variables to a Gamma log-link GLM, we discovered that two architecture-related variables — Parameters and Dataset Measurement — don’t exhibit a major relationship with useful resource consumption (p > 0.5). A multicollinearity check additionally confirmed that Dataset Measurement and Coaching Compute had been extremely correlated with different predictors (GVIF > 6). Based mostly on this, I hypothesized that each one three structure variables—Parameters, Dataset Measurement, and Coaching Compute) might not maintain a lot predictive energy. I then eliminated all three variables from the mannequin and an ANOVA check confirmed that simplified fashions (Fashions 4 and 5) will not be considerably worse than the total mannequin (Mannequin 1), with p > 0.05:

Mannequin 1: Energy_kWh ~ Parameters + Training_compute_FLOP + Training_dataset_size + 
    Training_time_hour + Hardware_quantity + Training_hardware + 
    0
Mannequin 2: Energy_kWh ~ Parameters + Training_compute_FLOP + Training_time_hour + 
    Hardware_quantity + Training_hardware
Mannequin 3: Energy_kWh ~ Parameters + Training_dataset_size + Training_time_hour + 
    Hardware_quantity + Training_hardware
Mannequin 4: Energy_kWh ~ Parameters + Training_time_hour + Hardware_quantity + 
    Training_hardware + 0
Mannequin 5: Energy_kWh ~ Training_time_hour + Hardware_quantity + Training_hardware + 
    0
  Resid. Df Resid. Dev Df Deviance Pr(>Chi)  
1        46     108.28                       
2        47     111.95 -1  -3.6700  0.07809 .
3        47     115.69  0  -3.7471           
4        48     116.09 -1  -0.3952  0.56314  
5        49     116.61 -1  -0.5228  0.50604 

Shifting on with Mannequin 5, I discovered that Coaching Time and {Hardware} Amount confirmed vital optimistic relationships with Power Consumption (GLM: coaching time, t = 9.70, p-value < 0.001; {hardware} amount, t = 6.89, p-value < 0.001). All {hardware} varieties had been additionally statistically vital (p-value < 0.001), indicating robust variation in power use throughout differing types. Detailed outcomes are introduced under:

glm(system = Energy_kWh ~ Training_time_hour + Hardware_quantity + 
    Training_hardware + 0, household = Gamma(hyperlink = "log"), knowledge = df)

Coefficients:
                                                Estimate Std. Error t worth Pr(>|t|)    
Training_time_hour                             1.351e-03  1.393e-04   9.697 5.54e-13 ***
Hardware_quantity                              3.749e-04  5.444e-05   6.886 9.95e-09 ***
Training_hardwareGoogle TPU v2                 7.213e+00  7.614e-01   9.474 1.17e-12 ***
Training_hardwareGoogle TPU v3                 1.060e+01  3.183e-01  33.310  < 2e-16 ***
Training_hardwareGoogle TPU v4                 1.064e+01  4.229e-01  25.155  < 2e-16 ***
Training_hardwareHuawei Ascend 910             1.021e+01  1.126e+00   9.068 4.67e-12 ***
Training_hardwareNVIDIA A100                   1.083e+01  3.224e-01  33.585  < 2e-16 ***
Training_hardwareNVIDIA A100 SXM4 40 GB        1.084e+01  5.810e-01  18.655  < 2e-16 ***
Training_hardwareNVIDIA A100 SXM4 80 GB        1.149e+01  5.754e-01  19.963  < 2e-16 ***
Training_hardwareNVIDIA GeForce GTX 285        3.065e+00  1.077e+00   2.846  0.00644 ** 
Training_hardwareNVIDIA GeForce GTX TITAN X    6.377e+00  7.614e-01   8.375 5.13e-11 ***
Training_hardwareNVIDIA GTX Titan Black        6.371e+00  1.079e+00   5.905 3.28e-07 ***
Training_hardwareNVIDIA H100 SXM5 80GB         1.149e+01  6.825e-01  16.830  < 2e-16 ***
Training_hardwareNVIDIA P100                   5.910e+00  7.066e-01   8.365 5.32e-11 ***
Training_hardwareNVIDIA Quadro P600            5.278e+00  1.081e+00   4.881 1.16e-05 ***
Training_hardwareNVIDIA Quadro RTX 4000        5.918e+00  1.085e+00   5.455 1.60e-06 ***
Training_hardwareNVIDIA Quadro RTX 5000        4.932e+00  1.081e+00   4.563 3.40e-05 ***
Training_hardwareNVIDIA Tesla K80              9.091e+00  7.760e-01  11.716 8.11e-16 ***
Training_hardwareNVIDIA Tesla V100 DGXS 32 GB  1.059e+01  6.546e-01  16.173  < 2e-16 ***
Training_hardwareNVIDIA Tesla V100S PCIe 32 GB 1.089e+01  1.078e+00  10.099 1.45e-13 ***
Training_hardwareNVIDIA V100                   9.683e+00  4.106e-01  23.584  < 2e-16 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for Gamma household taken to be 1.159293)

    Null deviance: 2.7045e+08  on 70  levels of freedom
Residual deviance: 1.1661e+02  on 49  levels of freedom
AIC: 1781.2

Variety of Fisher Scoring iterations: 25

Last Mannequin Choice

To raised seize potential non-additive results, varied interplay phrases had been explored and their respective AIC scores (Desk 1). The desk under summarizes the examined fashions and their respective AIC scores:

Though AIC scores didn’t range drastically, which means their mannequin matches are related, Mannequin 8 was most popular because it was the one one with vital results in each fundamental phrases and interplay. Interactions concerned {Hardware} Sort weren’t vital regardless of some exhibiting higher AIC, doubtless resulting from restricted pattern dimension throughout 18 {hardware} varieties.

In Mannequin 8, each Coaching Time and {Hardware} Amount confirmed a major optimistic relationship with power consumption (GLM: t = 11.09, p < 0.001), and between {hardware} amount and power consumption (GLM: coaching time, t = 11.09, p < 0.001; {hardware} amount, t = 7.32, p < 0.001; Fig. 3a). Their interplay time period was considerably detrimental (GLM: t = –4.32, p < 0.001), suggesting that power consumption grows extra slowly when coaching time will increase alongside with the next variety of {hardware} items. All {hardware} varieties remained vital (p < 0.001). Detailed outcomes are as under:

glm(system = Energy_kWh ~ Training_time_hour * Hardware_quantity + 
    Training_hardware + 0, household = Gamma(hyperlink = "log"), knowledge = df)

Coefficients:
                                                 Estimate Std. Error t worth Pr(>|t|)    
Training_time_hour                              1.818e-03  1.640e-04  11.088 7.74e-15 ***
Hardware_quantity                               7.373e-04  1.008e-04   7.315 2.42e-09 ***
Training_hardwareGoogle TPU v2                  7.136e+00  7.379e-01   9.670 7.51e-13 ***
Training_hardwareGoogle TPU v3                  1.004e+01  3.156e-01  31.808  < 2e-16 ***
Training_hardwareGoogle TPU v4                  1.014e+01  4.220e-01  24.035  < 2e-16 ***
Training_hardwareHuawei Ascend 910              9.231e+00  1.108e+00   8.331 6.98e-11 ***
Training_hardwareNVIDIA A100                    1.028e+01  3.301e-01  31.144  < 2e-16 ***
Training_hardwareNVIDIA A100 SXM4 40 GB         1.057e+01  5.635e-01  18.761  < 2e-16 ***
Training_hardwareNVIDIA A100 SXM4 80 GB         1.093e+01  5.751e-01  19.005  < 2e-16 ***
Training_hardwareNVIDIA GeForce GTX 285         3.042e+00  1.043e+00   2.916  0.00538 ** 
Training_hardwareNVIDIA GeForce GTX TITAN X     6.322e+00  7.379e-01   8.568 3.09e-11 ***
Training_hardwareNVIDIA GTX Titan Black         6.135e+00  1.047e+00   5.862 4.07e-07 ***
Training_hardwareNVIDIA H100 SXM5 80GB          1.115e+01  6.614e-01  16.865  < 2e-16 ***
Training_hardwareNVIDIA P100                    5.715e+00  6.864e-01   8.326 7.12e-11 ***
Training_hardwareNVIDIA Quadro P600             4.940e+00  1.050e+00   4.705 2.18e-05 ***
Training_hardwareNVIDIA Quadro RTX 4000         5.469e+00  1.055e+00   5.184 4.30e-06 ***
Training_hardwareNVIDIA Quadro RTX 5000         4.617e+00  1.049e+00   4.401 5.98e-05 ***
Training_hardwareNVIDIA Tesla K80               8.631e+00  7.587e-01  11.376 3.16e-15 ***
Training_hardwareNVIDIA Tesla V100 DGXS 32 GB   9.994e+00  6.920e-01  14.443  < 2e-16 ***
Training_hardwareNVIDIA Tesla V100S PCIe 32 GB  1.058e+01  1.047e+00  10.105 1.80e-13 ***
Training_hardwareNVIDIA V100                    9.208e+00  3.998e-01  23.030  < 2e-16 ***
Training_time_hour:Hardware_quantity           -2.651e-07  6.130e-08  -4.324 7.70e-05 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for Gamma household taken to be 1.088522)

    Null deviance: 2.7045e+08  on 70  levels of freedom
Residual deviance: 1.0593e+02  on 48  levels of freedom
AIC: 1775

Variety of Fisher Scoring iterations: 25

Coefficients Interpretation

To additional interpret the coefficients, we will exponentiate every coefficient and subtract one to estimate the p.c change within the response variable for every extra unit within the predictor (Popovic, 2022). For power consumption, every extra hour of coaching would enhance power use by 0.18%, every extra {hardware} unit would add 0.07%, and their interplay diminished their mixed fundamental results by 0.00002%. Equally, since water and carbon had been immediately proportional with power, the p.c change in coaching time, {hardware} amount, and their interplay remained the identical (Fig. 3b, Fig. 3c). Nonetheless, since {hardware} varieties had been categorical variables and functioned as baseline intercepts, their values differed throughout power, water, and carbon fashions to mirror variations in general scale.

RQ2: Power Effectivity over Time

I additionally used a log-linked Gamma mannequin to look at the connection between Power Effectivity and Publication Date, because the Shapiro-Wilk check indicated that the log-transformed knowledge was not usually distributed (p < 0.001). There was a optimistic relationship between Publication Date and Power Effectivity, with an estimated enchancment of 0.13% per 12 months (GLM: t = 8.005, p < 0.001, Fig. 3d).

To additional examine, we examined the developments by particular person {hardware} sort and noticed noticeable variation in effectivity amongst AI fashions utilizing the identical {hardware} (Fig. 3e). Amongst all structure and {hardware} decisions, Coaching Time was the one statistically vital issue influencing power effectivity (GLM: t = 8.581, p < 0.001), with longer coaching time decreases power effectivity by 0.03% per hour.

4. Dialogue

This research discovered that {hardware} decisions — together with {Hardware} Sort and {Hardware} Amount — together with Coaching Time, have a major relationship with every useful resource consumption throughout AI Model Training, whereas structure variables don’t. I think that Coaching Time might have implicitly captured a few of the underlying results of these architecture-related components. As well as, the interplay between Coaching Time and {Hardware} additionally contributes to the useful resource utilization. Nonetheless, this evaluation is constrained by the small dataset (70 legitimate fashions) throughout 18 {hardware} varieties, which doubtless limits the statistical energy of hardware-involved interplay phrases. Additional analysis may discover these interactions with bigger and extra various datasets.

For instance how resource-intensive AI coaching might be, we use Mannequin 8 to foretell the baseline power consumption for a single hour of coaching on one NVIDIA A100 chip. Listed here are the predictions for every sort of useful resource underneath this easy setup:

• Power: The expected power use is 29,213 kWh, almost thrice the annual power consumption of a median U.S. family (10,500 kWh/12 months) (U.S. Power Info Administration, 2023), with every further hour including 5258 kWh extra and every further chip including 2044 kWh.
• Water: Equally, the identical coaching session would eat 10,521 liters of water, nearly ten instances the common U.S. family’s every day water use (300 gallons or 1135 liters/day) (United States Environmental Safety Company, 2024), with every further hour including 1,894 liters and every chip including 736 liters.
• Carbon: the anticipated carbon emission is 16,009 kg, about 4 instances the annual emissions of a U.S. family (4000kg/12 months) (College of Michigan, 2024), with every further hour including 2881 kg and every further chip including 1120 kg.

This research additionally discovered that AI fashions have change into extra energy-efficient over time, however solely barely, with an estimated enchancment of 0.13% per 12 months. This means that whereas newer {hardware} is extra environment friendly, its adoption has not been widespread. Whereas the environmental influence of AI could also be mitigated over time as {hardware} {hardware} has change into extra environment friendly, this deal with {hardware} alone might overlook different contributors to general power consumption. On this dataset, each Coaching Compute and Whole Energy Draw are sometimes estimated values and should embrace some system-level overhead past {hardware} alone. Due to this fact, the effectivity estimates on this research might mirror not simply {hardware} efficiency, however probably different training-related overhead. This research noticed substantial variation in power effectivity even amongst fashions utilizing the identical {hardware}. One key discovering is that longer coaching time can “drain” power effectivity, decreasing it by roughly 0.03%. Additional research ought to discover how coaching practices, past {hardware} choice, influence the environmental prices of AI improvement.

References

Calvert, B.. 2024. AI already makes use of as a lot power as a small nation. It’s solely the start. Vox. https://www.vox.com/climate/2024/3/28/24111721/climate-ai-tech-energy-demand-rising

OpenAI Newsroom. 2024. Contemporary numbers shared by @sama earlier at this time: 300M weekly lively ChatGPT customers. 1B person messages despatched on ChatGPT on daily basis 1.3M devs have constructed on OpenAI within the US. Tweet by way of X. 2024. https://x.com/OpenAINewsroom/status/1864373399218475440

Epoch AI. 2025. Information on Notable AI Fashions. Epoch AI. https://epoch.ai/data/notable-ai-models

Shehabi, A., S.J. Smith, A. Hubbard, A. Newkirk, N. Lei, M.A.B. Siddik, B. Holecek, J. Koomey, E. Masanet, and D. Sartor. 2024. 2024 United States Information Middle Power Utilization Report. Lawrence Berkeley Nationwide Laboratory, Berkeley, California. LBNL-2001637.

Guidi, G., F. Dominici, J. Gilmour, Ok. Butler, E. Bell, S. Delaney, and F.J. Bargagli-Stoffi. 2024. Environmental Burden of United States Information Facilities within the Synthetic Intelligence Period. arXiv abs/2411.09786.

Bali, S.. 2025. GPU Reminiscence Necessities for AI Efficiency. NVIDIA Developer. https://developer.nvidia.com/blog/gpu-memory-essentials-for-ai-performance/

Krashinsky, R., O. Giroux, S. Jones, N. Stam, and S. Ramaswamy. 2020. NVIDIA Ampere Structure In-Depth. NVIDIA Developer. https://developer.nvidia.com/blog/nvidia-ampere-architecture-in-depth/

HuggingFace. 2025. Efficiency Suggestions for Coaching on A number of GPUs. HuggingFace Documentation. https://huggingface.co/docs/transformers/en/perf_train_gpu_many

Popovic, G.. 2022. Deciphering GLMs. Environmental Computing. Surroundings Computing. https://environmentalcomputing.net/statistics/glms/interpret-glm-coeffs/

U.S. Power Info Administration. 2023. Use of Power Defined: Electrical energy Use in Houses. https://www.eia.gov/energyexplained/use-of-energy/electricity-use-in-homes.php

United States Environmental Safety Company. 2024. How We Use Water. https://www.epa.gov/watersense/how-we-use-water

Middle for Sustainable Techniques, College of Michigan. 2024. Carbon Footprint Factsheet. Pub. No. CSS09–05.