TIMBER & INK
Home About The Digital Hearth Contact

The Digital Hearth · The Investigation

A  T I M B E R  &  I N K  E S S A Y

The Hidden Footprint of "Invisible" Intelligence

AI, water, energy, and what it means to be a good neighbor in a warming world.

If you've ever typed a prompt into ChatGPT or watched an image bloom from a handful of words, it can feel like magic: clean, quiet, weightless.

But "digital" doesn't mean immaterial.

Under every AI interaction is a physical reality — data centers, servers, cooling systems, power plants, transmission lines, and supply chains. And as the world races to build more AI capability, the environmental conversation has sharpened into two very earthy questions. How much electricity does this take? How much water does this drink?

What follows is a walk through the most common claims, an honest look at what needs correcting, and then a survey of what OpenAI, Google, Microsoft, and the broader industry are doing to reduce the impact — and how soon improvements may realistically arrive.

In This Investigation

  1. The Cloud Has a Body
  2. Energy: Checking the Numbers
  3. Water: Checking the "Thirsty" Claims
  4. Other Costs: Materials, E-Waste, Communities
  5. What's Being Done
  6. A Realistic Timeline
  7. Stewardship Without Panic

Sources & Further Reading →

01

The Cloud Has a Body

Where the conversation has to start.

The Claim

"AI is harmful to the environment because of data centers."

Verdict

Broadly valid, but details matter.

MIT summarizes the core issue well: training and running large generative AI models requires substantial electricity, and significant water can be used to cool the hardware. Manufacturing and transporting specialized hardware adds indirect impact, too (MIT News).

The "cloud" has a body. It's steel and circuitry, coolant lines and backup generators, and the rivers and power plants that keep it all from overheating.

The important nuance: not all AI is equally expensive. A tiny model running on-device is a different creature than a frontier model served at global scale from hyperscale data centers. The footprint varies dramatically by where the data center is — grid cleanliness, water availability, climate, cooling design.

02

Energy: Checking the Numbers

Four common claims, four honest verdicts.

Claim A

"Data centers used 4% of total U.S. electricity in 2023."

Verdict

Close, but the best-cited number is 2024, not 2023.

Pew, drawing on IEA estimates, reports that U.S. data centers consumed 183 TWh in 2024, described as "more than 4%" of total U.S. electricity use (Pew Research Center).

Claim B

"It will jump to 7–12% in the next three years."

Verdict

Plausible range, but scenario-dependent.

Some projections place data centers into high single digits or low double digits later this decade, especially if AI buildout continues at current speed (Pew Research Center). The Timber & Ink rule: present these numbers as scenarios, not promises.

Claim C

"A generative AI query uses 4–5x the energy of a typical search."

Verdict

The "multiple times higher" idea is supported; exact multipliers vary.

Some commonly cited estimates compare a web search at roughly 0.3 Wh to an LLM request at roughly 2.9 Wh, implying a several-fold difference. These are averages; real usage depends on model size, response length, caching, and serving efficiency (Epoch AI).

Claim D

"Training one large AI model emits as much carbon as five American cars over their lifetimes."

Verdict

Often repeated, frequently misunderstood or overgeneralized.

This line traces back to a widely-cited 2019 UMass Amherst analysis, popularized in multiple summaries and then spread as a shorthand. But even Google has noted that "five cars" has been misapplied in later retellings, and results vary dramatically depending on grid mix, hardware, and method (Google Research).

The right way to say it is slower: impact varies — sometimes wildly — depending on the power source, the cooling design, and the efficiency of the serving stack.
03

Water: Checking the "Thirsty" Claims

Two claims, and the careful framing they deserve.

Claim A

"A single large data center can use up to 5 million gallons of water per day."

Verdict

Possible for large facilities, but not typical.

Water use varies greatly by cooling system, climate, and load. EESI summarizes that water use at scale is substantial and growing, and the footprint can include both direct cooling and indirect water tied to electricity generation (EESI).

Claim B

"20–50 questions with ChatGPT can use about 500 ml of water."

Verdict

Based on an estimate, but highly context-dependent.

Researchers have explored the "water footprint" of AI workloads — cooling plus electricity — and some headlines translate this into "a bottle of water" per short session. But the exact number depends heavily on location, season, cooling design, and what's counted as "water used" (arXiv).

Responsible framing. It is fair to say AI can be water-intensive and that per-session water impacts can be meaningful in certain contexts. It is not fair to treat a single half-liter number as universal law.

04

Other Costs: Materials, E-Waste, Communities

Directionally valid, harder to quantify cleanly.

As AI expands, demand rises for high-performance hardware and for the electrical and construction systems that support it. These have upstream impacts — mining, manufacturing, shipping — and downstream impacts in the form of e-waste.

At the community level, concerns include grid strain and household cost impacts tied to the buildout (Pew Research Center). These are not abstractions. They show up on real bills, in real towns, with real people on either side of the meter.

05

What's Being Done

Where the major players say they're moving — and what actually moves the needle.

Microsoft

Cooling redesigns and water intensity targets

  • Microsoft has described a next-generation data center design that consumes zero water for cooling through chip-level cooling, rolling out beginning in 2024 (Microsoft).
  • Microsoft also describes a commitment to improve data center water-use intensity by 2030, and emphasizes closed-loop liquid cooling systems that recirculate (Microsoft On the Issues).

Google

Water replenishment and reporting

  • Google's published stewardship portfolio supports its goal to replenish 120% of freshwater consumed (on average) by 2030 (Google Sustainability).
  • Google also publishes environmental reporting that tracks energy and sustainability themes (Google Environmental Report).

OpenAI

Early public commitments and ongoing calls for transparency

OpenAI has begun addressing public concerns around data center strain and the need to avoid burdening communities, emphasizing what they describe as a "good neighbor" posture (The Verge).

The question isn't whether we should build. It's whether we will build like neighbors.

What Actually Moves the Needle

Beneath the corporate commitments, four practical levers determine whether AI's footprint shrinks meaningfully or just shrinks rhetorically.

i
Cleaner power. Emissions fall as the grid and corporate procurement get cleaner.
ii
Better cooling. Closed-loop and waterless designs reduce freshwater pressure.
iii
More efficient inference. Model and serving optimizations cut cost per request.
iv
Smarter siting. Climate and water conditions matter — deeply.
06

A Realistic Timeline

When this is likely to get better — and where the real bend in the curve sits.

Now — 2027
Demand growth may outpace efficiency gains in many regions. The footprint grows even as efficiency improves, because total volume is rising faster than per-unit savings.
2027 — 2030
Major public targets land here. Water stewardship commitments and new cooling designs begin reaching scale. The pace of efficiency gains catches up — or doesn't.
Post-2030
If clean power, advanced cooling, and efficient inference all scale together, per-unit impact can fall meaningfully. The variables that determine that future are being decided now.
07

Stewardship Without Panic

A word, especially, to my fellow Christians.

If you're a Christian using AI, the goal isn't fear or denial — it's wisdom. The same instinct that asks where does my food come from or who made this shirt applies here. A tool isn't morally neutral just because it's invisible. The question is whether we use it the way good neighbors use power.

Five Rails for Walking Upright

i
Tell the truth. Don't share inflated numbers because they're persuasive. Cite sources. Admit uncertainty.
ii
Love your neighbor, not just convenience. Water stress and grid strain are moral categories because people live downstream.
iii
Use AI like a tool, not an excuse. No plagiarism. No counterfeit trust.
iv
Practice digital temperance. Fewer prompts. Fewer regenerations. More intention.
v
Advocate for transparency. Innovation isn't the enemy; opacity is.
AI isn't pure evil, and it isn't pure miracle. It's power — and power always casts a shadow.

Will we be faithful with what we've been given?

For Further Reading

Sources cited above, in order of appearance.

  • MIT News — Explained: Generative AI's Environmental Impact
  • Pew Research Center — Energy Use at U.S. Data Centers Amid the AI Boom
  • EESI — Data Centers and Water Consumption
  • Google Research — Good News About the Carbon Footprint of Machine Learning Training
  • Epoch AI — How Much Energy Does ChatGPT Use?
  • arXiv — Making AI Less Thirsty (Water Footprint Research)
  • Microsoft — Zero Water for Cooling Data Centers
  • Microsoft On the Issues — Community-First AI Infrastructure
  • Google Sustainability — Water Stewardship Portfolio
  • Google — 2025 Environmental Report
  • The Verge — OpenAI Data Center Opposition & Energy Bills

Prefer a printable version to mark up by hand?

Download as PDF
Return to The Digital Hearth
—  ✦  —
Now go and make something of your spare moments. — J.
0
— ✦ —

Written by Joshua Davis from Tulsa, Oklahoma — a hospital tech worker, student minister, and Eagle Scout who believes ordinary life deserves extraordinary attention. Timber & Ink is where he writes about slow hobbies, analog craft, and the sacred in the ordinary.

An occasional letter.

Notes from Tulsa on paper, pens, tea, and whatever slow project is taking my time lately. That's all.

More slow stories

From the journal Browse all posts From the journal Browse all posts From the journal Browse all posts

Now go and make something of your spare moments.

— J.