Sources & Method

Sources and Method

This page shows the current calculator values, the sources behind them, and the method notes used to build the totals. Use it to inspect the evidence, assumptions, and system boundaries behind each comparison.

Calculator methods

Live row values

Calculator activity reference

Activity	Unit	Status	Cloud/source-side estimate	Broader-system estimate	Direct water	Total water	Sources
AI text promptsUses an everyday text-prompt estimate of about 0.3 Wh while keeping a separate prompt-specific water estimate of 0.26 mL direct and 1.3 mL total.	prompts Basic text questions	Estimated	0.3 Wh	0.3 Wh	0.26 mL	1.3 mL	Epoch AI (2025)Jegham et al. (2025)
High-reasoning promptsUses the long high-reasoning 33.8 Wh benchmark rather than the much smaller medium-query comparison value, so this row stays an intentional outlier.	prompts Long, high-reasoning queries	Inferred	33.8 Wh	33.8 Wh	33.8 mL	202.8 mL	Jegham et al. (2025)Internal synthesis (2026)
AI image generationUses the IEA's 1.7 Wh SD-XL 1.0-base benchmark from controlled H100 testing. Jon Ippolito's What Uses More? comparison prompted the review; the adopted value comes from the underlying IEA report. The benchmark covers GPU electricity only and is carried through both energy columns as a conservative floor. Water remains an internal conversion rather than an image-specific measurement.	images One SD-XL 1.0-base image under controlled H100 test conditions	Inferred	1.7 Wh GPU electricity only; excludes network energy, facility overhead, and the end-user device.	1.7 Wh Carries the 1.7 Wh GPU-only benchmark through as a conservative floor; this is not a measured total-system value.	1.7 mL	10.2 mL	Cozzi et al. (2025)Ippolito (2025)Internal synthesis (2026)
Coding-agent useUses the current one-hour coding-agent benchmark and adds a 30 Wh device allowance in the total-system column.	hours Hours of sustained agent use, not prompts	Estimated	325 Wh	355 Wh	325 mL	1950 mL	Simon P. Couch (2026)Internal synthesis (2026)
TikTok or short-video scrollingBuilt from device-side social-video measurement plus streaming-style infrastructure assumptions; keep this row estimated until a platform-specific cloud-side benchmark exists.	hours Video-heavy social feed	Estimated	7.5 Wh	36 Wh	7.5 mL	45 mL	Greenspector (2021)Adam Holter (2025)Kamiya (2020)Internal synthesis (2026)
Instagram scrollingThis remains an analogy-based estimate built from social-media measurement plus streaming comparisons, useful for relative scale rather than platform accounting.	hours Mixed image and short-video feed	Estimated	5 Wh	25 Wh	5 mL	30 mL	Greenspector (2021)Kamiya (2020)Internal synthesis (2026)
SnapchatAnother low-confidence social-media estimate built from broader comparison logic rather than a platform disclosure.	hours Short-form messaging and video	Estimated	4 Wh	20 Wh	4 mL	24 mL	Greenspector (2021)Internal synthesis (2026)
YouTube watchingUses the same CDN-style video benchmark as Netflix so the calculator keeps an apples-to-apples streaming anchor for the cloud-side and total-system columns.	hours Streaming video hour	Inferred	22 Wh	26 Wh	22 mL	132 mL	Kamiya (2020)Internal synthesis (2026)
Netflix streamingUses the IEA/Kamiya 22 Wh cloud-side and 77 Wh total benchmark used throughout the site's streaming comparisons.	hours One hour streamed	Inferred	22 Wh	37 Wh	22 mL	132 mL	Kamiya (2020)Internal synthesis (2026)
Email, cloud docs, and browsingThis is a teaching bundle for light browsing, email, and cloud documents rather than a per-click benchmark.	daily blocks A light mixed-work block	Estimated	4 Wh	30 Wh	4 mL	24 mL	Internal synthesis (2026)
Zoom as participantUses a one-hour participant meeting estimate in which network plus platform routing form the cloud-side figure and participant device energy is added back for the total-system column.	hours Joining a class or meeting	Estimated	54 Wh	74 Wh	54 mL	324 mL	David Mytton (2023)Verdecchia et al. (2022)Internal synthesis (2026)
Zoom as hostUses a hosted Zoom estimate that counts all participant devices in the total-system column.	meetings One 50-minute, 10-person hosted session	Estimated	192 Wh	358 Wh	192 mL	1152 mL	David Mytton (2023)Verdecchia et al. (2022)Internal synthesis (2026)
Gaming Console (Xbox Series X + TV)Locally rendered online multiplayer, not cloud gaming. The central estimate includes the console, reference TV, allocated home networking, fixed access and core transport, modeled game-server IT, and data-center overhead. Downloads, patches, manufacturing, standby, and optional texture streaming are excluded.	hours One active player-hour; local rendering over fixed broadband	Estimated	7.7 Wh Off-prem fixed access and core transport plus modeled game-server facility electricity; excludes the allocated home gateway.	220 Wh Console, reference TV, allocated home gateway, fixed access and core transport, modeled game-server IT, and PUE overhead for one player-hour.	0.32 mL Modeled direct cooling and humidification water for the game-server data center only; telecom-facility water is unquantified.	960 mL U.S.-average operational water consumption: electricity-generation water for quantified components plus modeled direct game-server data-center water.	Microsoft Xbox (2026)Microsoft Xbox (2026)Activision Publishing (2024)Mytton et al. (2024)U.S. Environmental Protection Agency (2024)Lyu et al. (2024)Microsoft PlayFab (2026)Lawrence Berkeley National Laboratory (2024)Internal synthesis (2026)

Source list

Calculator sources

This table lists the main sources used in the calculator and the role each one plays in the comparison model.

Source	Used for	Notes
How Much Energy Does ChatGPT Use?Epoch AI (2025) · Tier 3 · Expert analysis	0.3 Wh AI text-prompt row prompt-specific comparison language in the boundary and water method sections	Used for the 0.3 Wh everyday text-prompt estimate and related prompt-volume framing.
How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM InferenceJegham et al. (2025) · Tier 1 · Peer-reviewed or preprint benchmark	33.8 Wh high-reasoning prompt row direct and total water framing for prompt-heavy rows maintenance note preserving the long high-reasoning benchmark	Primary inference benchmark for reasoning-heavy prompt classes and environmental multipliers.
Digital inventory calculator and scenario methodInternal synthesis (2026) · Tier 3 · Internal synthesis	scenario-style 1 mL direct to 6 mL total water bridge browsing bundle row preset mix logic total-system add-ons for non-AI rows	Site-built method model that combines prompt benchmarks, streaming comparisons, meeting estimates, and water-conversion rules into the live calculator.
AI, Data Centers and Energy Demand: Reassessing and Exploring the TrendsRoucy-Rochegonde et al. (2025) · Tier 2 · Policy report	historical comparison for the pre-v0.2 image-generation row comparison logic in the cloud-side versus total-system bridge	Historical comparator for the pre-v0.2 image coefficient and useful for broader Jevons framing; it is not the source of the current image value and should be paired with Tier 1 sources.
Electricity use of AI coding agentsSimon P. Couch (2026) · Tier 3 · Expert analysis	coding-agent use row Gen Z vibe-coder preset cloud-side versus total-system bridge for high-intensity AI use	Useful for the vibe-coding scenario but should remain marked as estimated and napkin-math-derived.
What is the environmental footprint for social media applications? 2021 EditionGreenspector (2021) · Tier 2 · Industry measurement	TikTok, Instagram, and Snapchat directional rows cloud-side versus total-system bridge for social-video estimates maintenance note on low-confidence social rows	Used for device-side social-media energy measurements, especially TikTok.
Why Your ChatGPT Prompt Uses Half the Energy of a TikTok VideoAdam Holter (2025) · Tier 4 · Opinion or blog analysis	TikTok total-system estimate cross-check	Directional comparison used only as a low-confidence reference for TikTok total-system estimates.
The carbon footprint of streaming video: fact-checking the headlinesKamiya (2020) · Tier 1 · Institutional commentary	Netflix row YouTube row streaming-side values used to anchor TikTok and Instagram estimates	Used for the 77 Wh total streaming baseline and the ~22 Wh server-plus-network split adopted for Netflix and YouTube comparisons.
Zoom, video conferencing, energy, and emissionsDavid Mytton (2023) · Tier 3 · Expert analysis	Zoom participant row Zoom host row hosted Zoom derivation table worker-hosted-Zoom preset	Used for Zoom scenario modeling; should stay labeled as an estimate rather than a direct measurement of every meeting type.
Video Calls in Mobile Applications: Energy Consumption and Performance AnalysisVerdecchia et al. (2022) · Tier 1 · Peer-reviewed paper	participant device add-on for Zoom total-system device add-on in the hosted Zoom derivation	Used for device-side video-call energy and camera-on versus camera-off impacts in the Zoom estimates.
Measuring the environmental impact of delivering AI at Google ScaleSchneider (2025) · Tier 1 · Research paper	water methodology context on direct cooling versus broader electricity water	Primary source for WUE framing, cooling tradeoffs, and provider-scoped production prompt metrics.
2024 United States Data Center Energy Usage ReportLawrence Berkeley National Laboratory (2024) · Tier 1 · National laboratory report	water methodology context on electricity-side water intensity gaming-console PUE, WUE, and operational-water scenario	Primary U.S. demand, PUE/WUE, and electricity-related water-intensity context. Gaming uses its U.S. hyperscale and grid-water values only as a modeled scenario, not as provider-specific evidence.
What Uses More? AI Impact Calculator: sourcesIppolito (2025) · Tier 3 · Secondary calculator synthesis	provenance for the image-generation evidence review external comparison for the generic water-conversion method	External comparison and provenance source. Ippolito's spreadsheet prompted the image and water review, but underlying studies remain the evidentiary authority for adopted values.
Energy and AICozzi et al. (2025) · Tier 1 · International agency report	global electricity-demand and infrastructure context 1.7 Wh per image benchmark for SD-XL 1.0-base under controlled H100 test conditions	Primary institutional source for the adopted 1.7 Wh image-generation benchmark; the reported boundary is controlled H100 testing and GPU electricity only. Also used for global electricity-demand context.
The ML.ENERGY Leaderboard v3.0Chung et al. (2026) · Tier 1 · Academic benchmark and dataset	comparative image-generation evidence and configuration sensitivity	Academic benchmark used as comparative evidence for how model, steps, hardware, batching, and resolution change image-generation energy. Current leaderboard results must not be substituted silently for historical configurations cited by other calculators.
We did the math on AI's energy footprint. Here's the story you haven't heardO'Donnell et al. (2025) · Tier 2 · Technical journalism and analysis	comparative image-generation estimate in the Ippolito crosswalk	Secondary technical analysis used as a comparison point, including an image-generation estimate derived from measured joules. It does not replace a pinned primary benchmark.
Our contribution to a global environmental standard for AIMistral AI (2025) · Tier 2 · Industry environmental assessment	upper-bound context in the generic water-conversion review	Provider environmental assessment reporting 45 mL water and 1.14 gCO2e for a 400-token Le Chat response. It is useful for bounding water estimates but does not directly report mL per Wh.
Xbox Sustainability platform baselinesMicrosoft Xbox (2026) · Tier 1 · First-party laboratory measurement	Xbox Series X multiplayer-active high scenario gaming-console method	First-party laboratory platform baseline measured at the wall socket. The 150 W Series X multiplayer-active figure is used as the high console-power scenario, not as a Call of Duty-specific measurement.
Introduction to global platform baselinesMicrosoft Xbox (2026) · Tier 1 · First-party global telemetry model	Xbox Series X central console coefficient gaming-console method	First-party AC-wall telemetry model. The 127 W Series X shooter-genre average is the central console coefficient for the fixed gaming reference row.
Call of Duty Matchmaking Series: PingActivision Publishing (2024) · Tier 1 · First-party technical white paper	gaming functional unit and server architecture gaming-console method	First-party architectural evidence for dedicated servers, 10 to 12-player Core Multiplayer, distributed data centers, and reserved idle capacity. It reports no server electricity coefficient.
Network energy use not directly proportional to data volumeMytton et al. (2024) · Tier 1 · Peer-reviewed paper	fixed-broadband network component model gaming-console method	Peer-reviewed power model used to separate fixed access, core-network, and home CPE allocations from the much smaller usage-sensitive term.
ENERGY STAR Certified TelevisionsU.S. Environmental Protection Agency (2024) · Tier 1 · Government certification dataset	79.5 W reference television coefficient gaming-console method	Official certification dataset. The fixed reference uses the Sansui LE-55TA1 record: 54.7-inch screen and 79.5 W average on-mode power for certification.
Network Anatomy and Real-Time Measurement of Nvidia GeForce NOW Cloud GamingLyu et al. (2024) · Tier 1 · Peer-reviewed conference paper	central 0.20 Mbps gaming-traffic assumption gaming-console method	Peer-reviewed cloud-gaming study used only for its contextual statement that a typical locally rendered console game requires roughly 100 to 200 kbps. It is not a Call of Duty packet trace.
PlayFab Multiplayer Server termsMicrosoft PlayFab (2026) · Tier 2 · First-party platform documentation	generic multiplayer-server architecture analogy gaming-console method	Generic game-server architecture showing multiple containers per VM and standby capacity. It is analogy evidence only; no public evidence establishes that Call of Duty uses PlayFab or these configurations.

Cloud-side versus total-system bridge

The calculator keeps the first energy column as a cloud-side comparison view and then adds a second total-system column where the research basis supports it.

The cloud-side column usually combines server and network demand where the underlying source does not isolate pure data-center energy.
AI text and reasoning rows currently keep cloud-side and broader-system energy equal because the live calculator does not model a separate end-user device add-on for those actions.
The image row is narrower: its 1.7 Wh source covers GPU electricity only. That value is carried through both energy columns as a conservative floor and must not be described as measured server-plus-network or total-system energy.
Media and Zoom rows use a larger bridge from cloud-side to total-system values, which is why those rows show a bigger gap between columns.

Current bridge rules used in the calculator

Row family	Cloud-side basis	Total-system basis	Current calculator result	Maintenance note
AI text / reasoning prompts	Inference benchmarks used for everyday prompting	Kept equal to cloud-side in the live calculator	0.3/0.3 and 33.8/33.8	No separate end-user device model is currently applied
AI image generation	1.7 Wh GPU-only IEA benchmark	No boundary-matched broader-system benchmark; GPU-only value carried through as a conservative floor	1.7/1.7 Wh per image	Do not label either stored value as a measured server-plus-network or total-system result
Coding-agent use	High-intensity coding-agent benchmark	Adds a 30 Wh device allowance	325/355 Wh per hour	Keep this tied to the current coding-agent estimate, not to prompt counts
Netflix / YouTube	22 Wh per hour server + network	77 Wh per hour total-system benchmark	22/77 Wh per hour	Uses the IEA/Kamiya streaming split already adopted on the site
TikTok	Estimated 7.5 Wh per hour cloud-side	Estimated 36 Wh per hour total-system	7.5/36 Wh per hour	Built from social-video measurement plus streaming-style inference
Instagram	Estimated 5 Wh per hour cloud-side	Estimated 25 Wh per hour total-system	5/25 Wh per hour	Analogy-based row; keep estimated status
Snapchat	Estimated 4 Wh per hour cloud-side	Estimated 20 Wh per hour total-system	4/20 Wh per hour	Lowest-confidence social-media row in the calculator
Email / cloud docs / browsing	Internal light-work bundle	Bundle with device add-on	4/30 Wh per daily block	Not a per-click measurement
Zoom as participant	One-hour participant estimate	Adds participant device energy	54/74 Wh per hour	One-hour meeting framing
Zoom as host	Hosted meeting derivation used on the site	Adds all participant devices	192/358 Wh per meeting	One 50-minute hosted session with 10 people
Gaming Console (Xbox Series X + TV)	7.7 Wh off-prem fixed access, core transport, and modeled game-server facility electricity	Adds the Series X, reference 54.7-inch TV, and allocated home gateway	7.7/220 Wh per player-hour	Locally rendered multiplayer, not cloud gaming; infrastructure is attributional and the server term is low-confidence

SourcesEpoch AI (2025)Jegham et al. (2025)Cozzi et al. (2025)Microsoft Xbox (2026)Mytton et al. (2024)Ippolito (2025)Roucy-Rochegonde et al. (2025)Kamiya (2020)Greenspector (2021)David Mytton (2023)Verdecchia et al. (2022)Simon P. Couch (2026)Internal synthesis (2026)

Water methodology carried into the calculator

The calculator keeps prompt-specific direct and total water figures where they exist, then uses a broader electricity-to-water bridge for most of the remaining rows.

The current model rounds direct cooling water to roughly 1 mL per Wh of server energy and total water to roughly 6 mL per Wh once indirect electricity water is added back.
Text prompts keep their own prompt-specific direct and total water values instead of using the generic 1-to-6 rule.
Reasoning prompts, image generation, social video, streaming, browsing, and Zoom rows currently follow the broader proportional water rule unless a stronger prompt-specific figure exists.
Jon Ippolito's What Uses More? spreadsheet prompted a review of the generic coefficient. Its 7 mL/Wh selection is retained here as comparative synthesis, not adopted as a measured value.
Using the spreadsheet's stated 0.3 gCO2/Wh factor, Mistral's 45 mL and 1.14 gCO2e figures imply 11.84 mL/Wh, not 13 mL/Wh. The current 6 mL/Wh broader-water rule therefore remains unchanged pending a boundary-matched source.

Evidence considered in the 2026 water-method review

Evidence or method	Water relationship	Boundary	Current decision
Google production Gemini prompt	0.26 mL / 0.24 Wh = 1.08 mL/Wh	Direct data-center cooling water for Google's production serving boundary	Supports the rounded 1 mL/Wh direct-water bridge
Mistral 400-token Le Chat response	45 mL water and 1.14 gCO2e; not reported per Wh	Provider life-cycle assessment for one response	Upper-bound context only; do not treat as a direct mL/Wh measurement
Ippolito What Uses More? synthesis	7 mL/Wh, selected as a midpoint between about 1 and 13	Universal calculator conversion	Recorded as external comparison; not adopted
Recalculation of Ippolito's Mistral bridge	45 / 1.14 x 0.3 = 11.84 mL/Wh	Cross-metric conversion through a carbon-intensity assumption	Documents the arithmetic correction; still not boundary-matched evidence
Your Digital Life generic bridge	1 mL/Wh direct; 6 mL/Wh broader	Applied to the calculator's row-specific cloud-side energy field	Retained for v0.2.0 pending stronger activity-specific evidence

Water conversion rules used by the live calculator

Row family	Direct water	Total water	Basis	Maintenance note
Text prompts	0.26 mL per prompt	1.3 mL per prompt	Prompt-specific water estimate	Keep this special case unless the underlying prompt benchmark changes
Reasoning prompts	33.8 mL per prompt	202.8 mL per prompt	33.8 Wh energy row with a 1:6 water bridge	33.8 x 6 = 202.8
Image generation	1.7 mL per image	10.2 mL per image	1.7 Wh IEA image benchmark with a 1:6 water bridge	1.7 x 6 = 10.2
Social media / streaming / browsing rows	1 mL per Wh	6 mL per Wh	Rounded direct-plus-indirect water bridge	Used for TikTok, Instagram, Snapchat, YouTube, Netflix, and browsing
Zoom participant / host	1 mL per Wh	6 mL per Wh	Rounded direct-plus-indirect water bridge	Used for both meeting rows
Coding-agent use	325 mL per hour	1,950 mL per hour	1:6 water bridge	325 x 6 = 1,950
Gaming Console (Xbox Series X + TV)	0.32 mL per player-hour	960 mL per player-hour	Activity-specific LBNL U.S.-average operational-water scenario, not the generic 1:6 bridge	Includes electricity-generation water for console, TV, home gateway, access/core network, and server facility; direct telecom water remains unquantified

SourcesSchneider (2025)Lawrence Berkeley National Laboratory (2024)Epoch AI (2025)Jegham et al. (2025)Roucy-Rochegonde et al. (2025)Internal synthesis (2026)Ippolito (2025)Mistral AI (2025)Microsoft Xbox (2026)

Hosted Zoom meeting derivation

The calculator's hosted Zoom row combines a small routing/server term, a larger network term, and participant-device energy.

The server term is intentionally small because the model assumes SFU-style routing rather than heavy re-encoding.
The network term dominates the cloud-side figure and the participant devices dominate the total-system add-on.
This remains an estimate and should stay visibly separate from directly measured prompt benchmarks.

One hosted Zoom meeting with 10 participants over 50 minutes

Energy component	Per-meeting energy (Wh)	Basis
Zoom server routing	7	SFU-style routing estimate
Network across all participants	185	Mytton energy-per-GB framing
Devices across all participants	166	Laptop-energy assumption used for participant devices
Server + network subtotal	192	Cloud-side figure used in the calculator
Total-system energy	358	Hosted meeting figure used in the calculator

SourcesDavid Mytton (2023)Verdecchia et al. (2022)Internal synthesis (2026)

Image-generation evidence and Ippolito crosswalk

The image row now uses a pinned 1.7 Wh IEA benchmark discovered through comparison with Jon Ippolito's What Uses More? source sheet, while keeping Ippolito's 2 Wh figure and other published estimates visible as comparison evidence.

Jon Ippolito's What Uses More? spreadsheet prompted this review by selecting 2 Wh for a 1024x1024 image and exposing several candidate estimates.
The calculator does not treat Ippolito's 2 Wh synthesis as a primary measurement. It adopts the underlying IEA report's 1.7 Wh benchmark because the report identifies the model, H100 test hardware, controlled experimental context, and GPU-only boundary.
The IEA report does not establish a full production-serving or total-system value. The calculator keeps server and total energy equal for this row and labels the result inferred.
The image water figures are internal 1 mL/Wh direct and 6 mL/Wh broader conversions. None of the image-energy sources measures image-specific water use.

Image-generation evidence decision

Evidence	Figure	Workload and boundary	Decision role
International Energy Agency, Energy and AI, Figure 1.16 and accompanying text	1.7 Wh per image	SD-XL 1.0-base on H100 GPUs; controlled experimental conditions; GPU electricity only	Adopted v0.2.0 energy coefficient
Jon Ippolito, What Uses More? source sheet	2 Wh per 1024x1024 image	Secondary synthesis of multiple image estimates	Prompted the review and remains an explicit comparator; not treated as direct measurement
O'Donnell and Crownhart	1.2 Wh from 4,400 J	Stable Diffusion 3 with 50 denoising steps as summarized in Ippolito's sheet	Supporting comparison
ML.ENERGY historical result summarized by Ippolito	0.42 Wh from 1,500 J	Stable Diffusion 3 with 15 denoising steps; historical configuration must remain pinned	Lower comparison point; current leaderboard is versioned separately
Prior Your Digital Life coefficient	0.48 Wh per image	Exact model, hardware, dimensions, and derivation were not recoverable from the included Ifri source	Replaced in v0.2.0

SourcesCozzi et al. (2025)Ippolito (2025)Chung et al. (2026)O'Donnell et al. (2025)Roucy-Rochegonde et al. (2025)Internal synthesis (2026)

Gaming Console (Xbox Series X + TV) method

The fixed gaming row estimates one active player-hour of a locally rendered 10 to 12-player shooter on Xbox Series X with a reference 54.7-inch television and fixed broadband. It includes an attributional allocation of shared network infrastructure and a low-confidence modeled game-server term.

This is locally rendered online console play, not Xbox Cloud Gaming. Multiplayer traffic carries game state, voice, authentication, matchmaking, anti-cheat, and telemetry; the console renders the game.
The central calculator values are rounded because the infrastructure terms are modeled. Exact component arithmetic is retained below for reproducibility.
The low, central, and high cases are scenario bundles, not statistical confidence intervals.
The game-server IT coefficient is the weakest component. Activision publishes architectural details but no per-player electricity, host power, utilization, provider share, PUE, or WUE.

Central Xbox Series X plus TV component calculation

Component	Central energy	Boundary and evidence	Confidence
Xbox Series X console	127 Wh	Microsoft global shooter-genre AC-wall average	High for platform measurement; moderate for this workload
Reference television	79.5 Wh	EPA ENERGY STAR Sansui LE-55TA1, 54.7-inch average on-mode certification power; illustrative model, not a market average	High for the model; low for representativeness
Allocated home gateway / Wi-Fi	5.754 Wh	Mytton power model with 11.5 W idle allocated across two household users plus the 0.20% line-use term	Moderate-low
Fixed access network	5 Wh	Mytton central fixed-access allocation per subscriber	Moderate-low
Core / transit network	1.506 Wh	1.5 W baseline + 0.03 W per Mbps at the 0.20 Mbps central traffic assumption	Moderate-low
Game-server and support-service IT	1 Wh	Project synthesis allowance informed by dedicated-server architecture and generic containerized server hosting; not a Call of Duty measurement	Low
Data-center overhead	0.22 Wh	1 Wh server IT x (1.22 PUE - 1), using the LBNL hyperscale central scenario	Low
Exact quantified total	219.98 Wh	127 + 79.5 + 5.754 + 5 + 1.506 + 1 + 0.22; calculator rounds to 220 Wh	Moderate overall

Sensitivity scenarios for one active player-hour

Scenario	Console / display	Home + off-prem infrastructure	Total-system energy	Direct server-DC water	Broader operational water
Low	122 Wh / 23.65 Wh	6.497 Wh	152.147 Wh	0.010 mL	661.869 mL
Central	127 Wh / 79.5 Wh	13.480 Wh	219.980 Wh; rounded to 220	0.320 mL	957.440 mL; rounded to 960
High	150 Wh / 148.56 Wh	31.425 Wh	329.985 Wh	3.350 mL	1,440.213 mL

Published calculator fields and rounding

Published field	Calculation	Published value	Boundary
Cloud/source-side energy	5 access + 1.506 core + 1.22 server facility = 7.726 Wh	7.7 Wh	Off-prem access, core, and game-server facility; home gateway excluded
Broader-system energy	127 console + 79.5 TV + 5.754 home gateway + 7.726 off-prem = 219.98 Wh	220 Wh	Operational player-hour; manufacturing, downloads, patches, standby, and cloud rendering excluded
Direct water	1 Wh server IT x 0.32 L/kWh IT = 0.32 mL	0.32 mL	Modeled game-server data-center cooling and humidification only
Broader operational water	218.76 non-DC Wh x 4.35 mL/Wh + 1.22 DC Wh x 4.52 mL/Wh + 0.32 mL direct = 957.44 mL	960 mL	U.S.-average electricity-generation water consumption plus modeled direct game-server DC water

SourcesMicrosoft Xbox (2026)Microsoft Xbox (2026)Activision Publishing (2024)Mytton et al. (2024)U.S. Environmental Protection Agency (2024)Lyu et al. (2024)Microsoft PlayFab (2026)Lawrence Berkeley National Laboratory (2024)Internal synthesis (2026)

Broader site source registry

Tier 2

Company environmental report

2025 Environmental Report

Google

Tier 2

Industry reporting

A new Google data center triggers 1.9 GW of clean energy in Minnesota

Electrek

Tier 2

Reported analysis

AI Is Draining Water From Areas That Need It Most

Bloomberg

Tier 1

Market monitor report

Analysis of the 2025/2026 RPM Base Residual Auction: Part G (Revised)

Monitoring Analytics

Tier 1

Market monitor report

Analysis of the 2027/2028 RPM Base Residual Auction: Part A

Monitoring Analytics

Tier 1

Reported case study

As Google's Water Demands Grow, The Dalles Aims to Pull More from Mount Hood Forest

Oregon Public Broadcasting

Tier 4

Vision essay

Machines of Loving Grace

Dario Amodei

Tier 2

Industry reporting

Meta, Amazon, Google, and Microsoft Dominate Clean Energy Deals

Carbon Credits

Tier 2

Industry announcement

One year later, Crane Clean Energy Center still in the spotlight and ahead of schedule

Constellation Energy

Tier 1

Grid operator release

PJM auction procures 134,479 MW of generation resources

PJM Interconnection

Tier 2

Advocacy analysis

Powering Data Centers with Clean Energy Could Avoid Trillions in Climate and Health Costs

Union of Concerned Scientists

Tier 2

Policy analysis

Projected data center growth spurs PJM capacity prices by factor of 10

IEEFA

Tier 2

Company announcement with secondary verification

Sustainable by Design: Zero Water Cooling

Microsoft / Serverion summary

Tier 1

Federal energy data summary

U.S. electric power sector continues water efficiency gains

U.S. Energy Information Administration

Tier 1

Research paper

Video Killed the Energy Budget: Characterizing the Latency and Power Regimes of Open Text-to-Video Models

arXiv

Tier 2

Regulatory summary

Virginia's new data center electricity rate class

American Action Forum