Digital Habits, Water & Energy

Sources & Method

Sources and Method

This page explains how figures are selected, how statuses are assigned, and how the calculator turns research into comparable public estimates.

  • Every calculator row should map to a cited source or a clearly labeled site-built estimate.
  • Cloud-side and total-system figures answer different questions and should not be collapsed into one number.
  • Verified, inferred, estimated, and contested claims should be visibly different.
  • Method notes matter because water and energy figures often change with scope.
  • This calculator shares an early lineage with Jon Ippolito's What Uses More? but uses a different activity mix and method posture.
  • The live tool is built for daily inventory, not just pairwise comparison.

Core rules

  • Use published research where possible, and label site-built bridges or bundles when direct measurements are missing.
  • Keep the calculator’s figures, citations, and method notes aligned so the public values and their sources stay readable together.
  • Keep server-only and total-system figures separate.
  • Keep direct and indirect water separate.
  • Mark analogies and extrapolations as estimates, not findings.
  • Retire figures when the supporting evidence no longer holds.

Relationship to What Uses More?

This calculator was built in conversation with Jon Ippolito’s What Uses More?, which was an early blueprint for putting AI tasks beside familiar digital habits. The family resemblance is deliberate. Both tools try to cut down on bad comparisons by keeping AI inside a broader digital-activity frame.

Still, the two calculators are not doing the same job. Jon’s public app is a two-task comparator with factor switches. This site’s calculator is a daily-inventory tool that totals many activities at once and keeps source status, system boundary, and water scope visible on the page.

Where the source overlap sits

The direct overlap is real, but it is partial.

  • Both tools use the Epoch AI baseline for everyday AI text prompting.
  • Both keep Mytton-style Zoom energy reasoning in play for video meetings.
  • Both use streaming and short-video comparisons so AI is not treated as the only digital habit worth measuring.

The divergence is bigger than the overlap.

  • This calculator leans on Jegham for reasoning-heavy prompts, Verdecchia for device-side meeting energy, Schneider/Google plus LBNL for water framing, Greenspector for social-media device measurements, and an explicit internal scenario-method record for the bridge from cloud-side to total-system values.
  • Jon’s current source sheet and factor table bring in ML.ENERGY, Delavande, Greenly, Google 2009, Mulkey, Mistral, EPA and Lawrence carbon factors, Luccioni and Dauner reasoning multipliers, and several self-authored calculations for Netflix, Zoom, and TikTok.
  • The result is two tools that share a teaching instinct and a few benchmark families, but do not rest on the same live source base.

How the calculators differ

What Uses More? is built for pairwise comparison. Users pick two tasks, then adjust energy source, climate, prompt complexity, prompt count, quantity, and output units. That makes it strong for equivalence questions and classroom sensitivity tests where the point is to see how one factor changes the answer.

This site’s calculator takes a different angle on digital habits. It asks what a day looks like when prompts, streaming, social media, browsing, and meetings stack together. It also keeps four totals visible at once: server + network energy, total-system energy, direct water, and total water. The activity mix reflects that goal. Jon’s tool includes one-off tasks like AI code, AI video, Google search, cloud storage, and phone charging. This site emphasizes rows that help people inventory a day, such as browsing blocks, participant vs. host Zoom, high-reasoning prompts, and coding-agent use by the hour.

Tradeoffs

Jon’s model is more interactive around scenario variation. It is easier to ask what happens if the same task runs on coal-heavy power, in a warm climate, or with reasoning turned on. The tradeoff is that more of the methodological nuance sits in factor tables and rationale text, and the user is still reading one task pair at a time.

This site’s model is better for seeing how everyday digital habits accumulate across a full day. It is also more explicit about system boundaries because cloud-side versus total-system energy and direct versus total water stay separate on the page. The tradeoff is that more assumptions are fixed in place. Users cannot yet toggle electricity mix, climate, or prompt complexity inside the live tool, and several non-AI rows remain scenario-style estimates rather than platform disclosures.

For substantive claims, cite the linked studies or the upstream Section 1-4 source files, not just the calculator interface.

Calculator methods

How the calculator is built

The calculator combines prompt benchmarks, streaming and meeting comparisons, and explicit water-conversion rules into one consistent comparison tool.

The tables below show the live row values, the source records tied to each row, and the assumptions used to move from cloud-side estimates to broader total-system comparisons.

Live row values

Calculator activity reference

Activity Unit Status Server + network Total-system Direct water Total water Sources
AI text prompts Uses 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 prompts Uses 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 generation Uses an open-model image-generation average and the same water-conversion rule used across the broader comparison set.
images
One generated image
 Inferred 0.48 Wh 0.48 Wh 0.48 mL 2.88 mL
Roucy-Rochegonde et al. (2025) Internal synthesis (2026)
Coding-agent use Uses 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
Couch (2026) Internal synthesis (2026)
TikTok or short-video scrolling Built 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) Holter (2025) Kamiya (2020) Internal synthesis (2026)
Instagram scrolling This 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)
Snapchat Another 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 watching Uses 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 77 Wh 22 mL 132 mL
Kamiya (2020) Internal synthesis (2026)
Netflix streaming Uses 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 77 Wh 22 mL 132 mL
Kamiya (2020) Internal synthesis (2026)
Email, cloud docs, and browsing This 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 participant Uses 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
Mytton (2023) Verdecchia et al. (2022) Internal synthesis (2026)
Zoom as host Uses 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
Mytton (2023) Verdecchia et al. (2022) Internal synthesis (2026)

Inline citation catalog

Calculator source registry

These source records now live in the same markdown file as the calculator values, method tables, and row annotations. Update the citation here first, then revise any tied figures.

Source ID Source Used for Notes
epoch-chatgpt-energy
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.
jegham-how-hungry-is-ai
How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference Jegham 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.
scenario-methods
Digital inventory calculator and scenario method Internal 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.
ifri-ai-dc-energy
AI, Data Centers and Energy Demand: Reassessing and Exploring the Trends Roucy-Rochegonde et al. (2025) · Tier 2 · Policy report
  • image-generation benchmark row
  • comparison logic in the cloud-side versus total-system bridge
 Useful for prompt comparison tables and the Jevons framing, but should be paired with Tier 1 sources where possible.
simon-couch-coding-agent
Electricity use of AI coding agents Couch (2026) · Tier 3 · Expert analysis
  • coding-agent use row
  • coding-day 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.
greenspector-social-media-2021
What is the environmental footprint for social media applications? 2021 Edition Greenspector (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.
adam-holter-tiktok
Why Your ChatGPT Prompt Uses Half the Energy of a TikTok Video 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.
kamiya-streaming-video
The carbon footprint of streaming video: fact-checking the headlines Kamiya (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.
mytton-zoom
Zoom, video conferencing, energy, and emissions Mytton (2023) · Tier 3 · Expert analysis
  • Zoom participant row
  • Zoom host row
  • hosted Zoom derivation table
  • hosted-zoom preset
 Used for Zoom scenario modeling; should stay labeled as an estimate rather than a direct measurement of every meeting type.
verdecchia-video-call-energy
Video Calls in Mobile Applications: Energy Consumption and Performance Analysis Verdecchia 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.
schneider-google-scale
Measuring the environmental impact of delivering AI at Google Scale Schneider (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.
lbnl-data-center-report
2024 United States Data Center Energy Usage Report Lawrence Berkeley National Laboratory (2024) · Tier 1 · National laboratory report
  • water methodology context on electricity-side water intensity
 Primary U.S. demand and water-intensity context for the macro-risk pages.

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, reasoning, and image rows currently keep server and total-system energy equal because the live calculator does not model a separate end-user device add-on for those actions.
  • 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 familyCloud-side basisTotal-system basisCurrent calculator resultMaintenance note
AI text / reasoning / image promptsInference benchmarks used for everyday promptingKept equal to cloud-side in the live calculator0.3/0.3, 33.8/33.8, 0.48/0.48No separate end-user device model is currently applied
Coding-agent useHigh-intensity coding-agent benchmarkAdds a 30 Wh device allowance325/355 Wh per hourKeep this tied to the current coding-agent estimate, not to prompt counts
Netflix / YouTube22 Wh per hour server + network77 Wh per hour total-system benchmark22/77 Wh per hourUses the IEA/Kamiya streaming split already adopted on the site
TikTokEstimated 7.5 Wh per hour cloud-sideEstimated 36 Wh per hour total-system7.5/36 Wh per hourBuilt from social-video measurement plus streaming-style inference
InstagramEstimated 5 Wh per hour cloud-sideEstimated 25 Wh per hour total-system5/25 Wh per hourAnalogy-based row; keep estimated status
SnapchatEstimated 4 Wh per hour cloud-sideEstimated 20 Wh per hour total-system4/20 Wh per hourLowest-confidence social-media row in the calculator
Email / cloud docs / browsingInternal light-work bundleBundle with device add-on4/30 Wh per daily blockNot a per-click measurement
Zoom as participantOne-hour participant estimateAdds participant device energy54/74 Wh per hourOne-hour meeting framing
Zoom as hostHosted meeting derivation used on the siteAdds all participant devices192/358 Wh per meetingOne 50-minute hosted session with 10 people
Sources Epoch AI (2025) Jegham et al. (2025) Roucy-Rochegonde et al. (2025) Kamiya (2020) Greenspector (2021) Mytton (2023) Verdecchia et al. (2022) 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.

Water conversion rules used by the live calculator

Row familyDirect waterTotal waterBasisMaintenance note
Text prompts0.26 mL per prompt1.3 mL per promptPrompt-specific water estimateKeep this special case unless the underlying prompt benchmark changes
Reasoning prompts33.8 mL per prompt202.8 mL per prompt33.8 Wh energy row with a 1:6 water bridge33.8 x 6 = 202.8
Image generation0.48 mL per image2.88 mL per image0.48 Wh image benchmark with a 1:6 water bridge0.48 x 6 = 2.88
Social media / streaming / browsing rows1 mL per Wh6 mL per WhRounded direct-plus-indirect water bridgeUsed for TikTok, Instagram, Snapchat, YouTube, Netflix, and browsing
Zoom participant / host1 mL per Wh6 mL per WhRounded direct-plus-indirect water bridgeUsed for both meeting rows
Coding-agent use325 mL per hour1,950 mL per hour1:6 water bridge325 x 6 = 1,950
Sources Schneider (2025) Lawrence Berkeley National Laboratory (2024) Epoch AI (2025) Jegham et al. (2025) Roucy-Rochegonde et al. (2025) Internal synthesis (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 componentPer-meeting energy (Wh)Basis
Zoom server routing7SFU-style routing estimate
Network across all participants185Mytton energy-per-GB framing
Devices across all participants166Laptop-energy assumption used for participant devices
Server + network subtotal192Cloud-side figure used in the calculator
Total-system energy358Hosted meeting figure used in the calculator
Sources Mytton (2023) Verdecchia et al. (2022) Internal synthesis (2026)

Starter mix provenance

The presets are loading aids, not canonical public claims. They are intentionally close to recurring use patterns without trying to reproduce one exact day.

  • Streaming Day is a fast load of a video-heavy pattern without AI.
  • Moderate AI Day mirrors a mixed day where text prompts and images are present but not dominant.
  • Hosted Zoom Day isolates the burden of one hosted meeting.
  • Coding Day is deliberately lighter than the heaviest coding-agent example so users can adjust from a plausible starting point rather than from the most extreme case.

Preset intent

PresetIntended basisCurrent loaded valuesMaintenance note
Streaming DayVideo-heavy non-AI starter1.5 TikTok, 1.25 YouTube, 0.75 Instagram, 0.5 Snapchat, 1.5 Netflix, 1 Zoom participantUse for fast loading, not as a published total
Moderate AI DayMixed prompt-and-media starter25 text prompts, 3 images, 1 TikTok, 1 YouTube, 0.5 Instagram, 0.25 Snapchat, 1 Netflix, 1 Zoom participantKeeps a mixed AI pattern visible
Hosted Zoom DayMeeting-heavy starter1 hosted Zoom meeting, 1 Instagram, 1 Netflix, 1 browsing blockDesigned to foreground the meeting row
Coding DayHigh-intensity coding starter1 coding-agent hour, 1 reasoning prompt, 15 text prompts, 2 images, 0.25 TikTok, 0.5 YouTube, 0.75 NetflixIntentionally less loaded than the heaviest coding-agent example
Sources Internal synthesis (2026) Couch (2026) Mytton (2023)

Current synthesis choices to preserve or revisit deliberately

  • The reasoning row intentionally uses the 33.8 Wh long high-reasoning benchmark, not the much smaller medium-reasoning comparison value.
  • The browsing row is a teaching bundle, not a platform disclosure, and should not be rewritten as if it were a direct measurement.
  • The TikTok, Instagram, and Snapchat rows remain estimated because the underlying platform-specific cloud-side numbers are still absent.
  • If stronger prompt, streaming, or meeting research appears, update the supporting note here before changing the rendered calculator values.
Sources Jegham et al. (2025) Greenspector (2021) Internal synthesis (2026)

Broader site source registry