Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsiTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
There is no universal energy or water cost per AI prompt. A defensible estimate starts by defining the model, task, request size, hardware, utilization, idle capacity, host systems, data-center overhead, location, date and water boundary. If provider-side workload or facility data is missing, publish a scenario or range rather than a falsely precise number.
This guide shows how to calculate electricity use, convert it to a monetary cost, estimate on-site cooling water, and explain what your result leaves out. It also shows why published figures from Google and Microsoft differ: they describe different workloads and measurement boundaries, not a single industry average.
What an AI energy-and-water estimate actually measures
An AI service has several possible accounting boundaries. Measuring only an accelerator while it is actively calculating is not the same as measuring a production request. A production estimate may also include the server’s CPU and memory, machines waiting for work, redundant capacity, networking, storage and the facility overhead needed to remove heat.
Water requires the same discipline. “Water used” can mean water evaporated at the data center, water withdrawn and returned, water consumed by electricity generation, or water embedded in manufacturing hardware. Those quantities are not interchangeable. Put the boundary in the headline or first sentence of every result.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
Start by defining the workload
Separate training from inference
Training is a long-running job that processes a dataset repeatedly and may use thousands of accelerators. Inference is the serving of a trained model, one request at a time or in batches. Report them separately: a training estimate divided by the number of later prompts can be useful for a lifecycle study, but it is not the energy of serving a prompt.
Describe the task and modality
Record whether the workload is text, image, audio, video, embedding, retrieval, fine-tuning or another operation. A text completion and a video-generation request do not perform equivalent work, even when both are called “one request.” Include batch size, concurrency and any tool calls when they materially change computation.
Record input and output work
For text inference, capture input tokens and output tokens for each request, then report the median and a spread such as the interquartile range (IQR). Token counts are useful workload descriptors, but they do not determine energy by themselves: architecture, precision, batching, accelerator type, utilization and decoding method also matter.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCount requests over a stated period
Choose a period such as one day, month or year. Keep successful, retried and failed requests in separate counts if they consumed compute. State whether cached responses, moderation passes, retrieval and tool calls are included.
Choose an electricity boundary before doing arithmetic
| Boundary | Included | When it is useful | Main limitation |
|---|---|---|---|
| Accelerator active power | Power drawn by GPUs, TPUs or other accelerators while calculating | Comparing kernel or hardware efficiency | Omits host systems, idle capacity and facility overhead |
| IT energy | Accelerators plus host CPU, DRAM and other server components | Estimating the equipment serving a workload | May still omit cooling and other facility loads |
| Production serving energy | IT energy plus provisioned idle machines, utilization effects and serving infrastructure | Operational per-request reporting | Requires provider-specific telemetry or modeled assumptions |
| Facility energy | Production IT energy multiplied by an appropriate power usage effectiveness (PUE) factor, when PUE is not already included | Comparing a request with the electricity entering a data center | PUE varies by site, weather, season and reporting population |
Do not multiply a published facility-inclusive figure by PUE again. That double-counts overhead. Conversely, if your number covers only IT equipment, state that and apply PUE once to obtain a facility estimate.
Calculate electricity use step by step
- Collect a per-request energy value. Prefer measured provider telemetry. If unavailable, use a model that states accelerator type, utilization, host energy, idle capacity and PUE assumptions.
- Attach the boundary to the value. Write “accelerator-only,” “IT,” or “facility-inclusive” next to the number; never leave the unit unexplained.
- Convert watt-hours to kilowatt-hours. Use
kWh per request = Wh per request ÷ 1,000. - Scale to traffic. Use
total kWh = kWh per request × number of requests. If requests have different sizes, calculate each workload class separately and add the results. - Convert consumption to an electricity bill. Use
energy cost = total kWh × your electricity tariff. State the currency, tariff period, demand charges and whether taxes or renewable surcharges are included. A provider’s internal cost may not equal a household or public tariff. - Show uncertainty. Publish a range, IQR or low/base/high scenario whenever utilization, request mix or infrastructure factors are estimated.
Published examples that demonstrate the range
| Example | Reported result | Boundary and conditions | What it demonstrates |
|---|---|---|---|
| Google Gemini Apps text prompt | 0.24 Wh for the median prompt | Google’s comprehensive method; May 2025 data; includes active accelerator power, host CPU and DRAM, idle machines and data-center overhead | A provider-specific production estimate, not a universal prompt benchmark |
| Google Gemini Apps text prompt | 0.10 Wh for the median prompt | Google’s narrower method that excludes several components | The boundary alone can materially change the answer |
| Microsoft Research frontier-scale model | 0.34 Wh median per query; IQR 0.18–0.67 Wh | Bottom-up model for an H100 node under stated workload, utilization and PUE assumptions | A modeled result cannot be transferred to every model or provider |
| Microsoft Research test-time scaling | 4.32 Wh median per query | Scenario using 15 times more tokens than a typical query | More reasoning tokens can raise energy by an order of magnitude |
The Google and Microsoft figures use different systems, dates and assumptions. They should not be averaged into one “AI prompt” number.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
Worked electricity example
Suppose an internal service has 2,000,000 requests in a month and you have a measured facility-inclusive value of 0.30 Wh per request. The monthly consumption is 0.30 × 2,000,000 ÷ 1,000 = 600 kWh. At a hypothetical tariff of 0.15 currency units per kWh, the energy charge would be 90 currency units before demand charges, taxes or other contract terms. The tariff is an input you must replace with the rate relevant to your site; it cannot be inferred from a model’s token count.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →If the 0.30 Wh figure were IT-only and the matching PUE were 1.20, the facility estimate would instead be 0.30 × 1.20 = 0.36 Wh per request. Applying 1.20 to a figure that already includes facility overhead would be incorrect.
Estimate water with a defined WUE factor
Understand WUE
Water Usage Effectiveness (WUE) is commonly expressed as liters per kilowatt-hour. Microsoft defines WUE as annual liters used for humidification and cooling divided by annual IT kWh. Its FY25 global WUE is 0.27 L/kWh for data centers Microsoft fully owns and controls that operated for 12 months. FY25 covers July 1, 2024 through June 30, 2025. That is a reporting-population and period-specific factor, not a universal value for another provider or location.
Match the numerator and denominator
Use on-site water (L) = matching IT energy (kWh) × WUE (L/kWh). The energy in the denominator must have the same IT boundary as the WUE factor. Do not multiply facility-inclusive energy by a WUE factor calculated from IT-only energy unless you first reconcile the boundaries.
If a source reports water in milliliters, convert with 1,000 mL = 1 L. To estimate a monetary water cost, use water cost = liters ÷ 1,000 × local water-and-sewer price per cubic meter, while noting that many data centers use reclaimed water, fixed contracts or internal accounting rather than a retail tariff.
Google’s prompt-water method
Google’s paper estimates prompt water as (total energy per prompt − data-center overhead energy per prompt) × previous-calendar-year average freshwater WUE Category 2. It reports 1.15 L/kWh for both 2023 and 2024. Google defines the selected Category 2 metric as consumptive use—water input minus returned water—intended to represent water evaporated and unavailable for reuse.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
Using its 2024 fleet-average factors, Google reports 0.26 mL of water for the median Gemini Apps text prompt. That is an operational provider estimate for the stated prompt population and method; it is not a full lifecycle water footprint.
Worked on-site water example
If your matched IT energy is 600 kWh for a month and the applicable WUE is 0.27 L/kWh, on-site cooling and humidification water is 600 × 0.27 = 162 L. If you instead use a site-specific consumptive WUE of 1.15 L/kWh, the result is 600 × 1.15 = 690 L. The difference does not show that one calculation is wrong: the factors describe different facilities, definitions or reporting populations. Use the factor whose boundary and geography match your energy data.
Do not confuse on-site water with a full water footprint
Cooling water is only one part of an AI system’s water impact. Electricity generation can consume water, and manufacturing accelerators, servers and buildings adds embodied water. Weather, cooling design, local water stress, grid mix and the time of day can change the result.
The 2024 study Making AI Less “Thirsty” discusses spatial and temporal differences and notes that public runtime and embodied-water data remain limited. If you include indirect water, state the generation-water factor, manufacturing scope and geography separately. Never add an unspecified generic number to a provider’s on-site figure.
How to compare two estimates fairly
Before deciding that one model is “greener,” align the following dimensions:
- Task and modality: compare text with text, image with image, and equivalent quality targets.
- Work per request: align input/output tokens, generated pixels or frames, audio duration, batch size and tool calls.
- Hardware: record accelerator model, number of devices, precision and memory configuration.
- Utilization: include throughput, concurrency, batching and tail latency targets.
- Idle and reserved capacity: state whether waiting or failover machines are allocated to requests.
- Host and facility energy: identify CPU, DRAM, networking, storage and PUE treatment.
- Water scope: distinguish on-site operational WUE from electricity-generation and embodied water.
- Date and geography: fleet averages can change as hardware, weather, grid mix and cooling systems change.
- Statistical summary: use the same percentile or median, and show IQR or another spread when available.
Google explains that its prompt distribution is skewed and says, “Consequently, we use the daily median as the aggregate value for the metrics defined in Section 3.2 and Section 3.3.” Treat that as the paper’s chosen reporting method, not a rule that every workload must use a median.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
Build a range when provider data is missing
Most commercial services do not expose request-level runtime, host energy, facility overhead, exact cooling water or embodied impacts. In that case, create scenarios rather than inventing precision.
| Scenario | Typical assumptions to vary | How to label it |
|---|---|---|
| Low | Short outputs, efficient accelerator, high utilization, little idle capacity and a favorable PUE | Illustrative lower-bound scenario, not a measured minimum |
| Base | Observed token distribution, expected utilization and a site or provider average | Best planning estimate under stated assumptions |
| High | Long outputs or reasoning, lower utilization, reserved capacity, retries and less favorable facility conditions | Planning upper scenario, not a worst-case physical limit |
Keep uncertain inputs visible. For example, if you know request count but not Wh per request, publish the assumed Wh range and show the resulting kWh range. If you know IT energy but not PUE, report IT kWh and a separate facility scenario instead of silently choosing a factor.
A practical worksheet
| Field | What to enter |
|---|---|
| Model and version | Exact model, deployment or API version and measurement date |
| Task | Inference, training, fine-tuning, embedding or another operation |
| Modality | Text, image, audio, video or mixed workflow |
| Workload size | Input/output tokens or equivalent task units; median and spread |
| Request volume | Requests in the reporting period, including retry policy |
| Per-request energy | Wh value, measurement method and boundary |
| Facility factor | PUE value and whether it is already included |
| Water factor | WUE in L/kWh, definition, site, period and consumptive/withdrawal status |
| Indirect scope | Electricity-generation water, manufacturing or neither |
| Prices | Electricity tariff and water/sewer rate, currency and billing period |
| Uncertainty | IQR, low/base/high scenarios or confidence range |
Common mistakes and how to correct them
Using tokens as an energy meter
Tokens describe part of the workload, not the hardware or facility. Correct this by pairing token counts with measured or modeled Wh and stating utilization and serving assumptions.
Reporting accelerator power as total service energy
Active accelerator readings omit CPU, memory, idle machines and cooling. Correct this by labeling the narrow result and adding host and facility components when the question concerns production service.
Applying PUE twice
Read the definition of the published energy number. Multiply by PUE only when the number is IT energy and the chosen PUE matches the site and period.
Calling every water number “water consumption”
Withdrawal, discharge, evaporation and consumptive use answer different questions. Name the metric, units, numerator, denominator and whether water is returned.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
Mixing fleet factors with request energy from another boundary
A fleet-average WUE cannot automatically be combined with an accelerator-only Wh value. Match IT-versus-facility boundaries, site population and time period first.
Turning one provider’s result into an industry average
Google’s Gemini figure and Microsoft’s H100 model illustrate specific systems. Use them as examples or scenario anchors, not as a universal conversion factor.
Hiding uncertainty behind many decimal places
Round to a level supported by the inputs and show a range when request mix, hardware or facility data is estimated. False precision is less useful than a transparent interval.
Free tools Windows power users keep installed
One-click scans. No signup required.
How to present a final result
A reader should be able to reproduce the calculation from one short statement and an accompanying table. Use this order:
- State the workload, model, date and geography.
- State the energy boundary and per-request value.
- Show request volume and the kWh conversion.
- State PUE treatment and calculate facility energy if needed.
- State WUE definition, site and period, then calculate on-site liters.
- List excluded impacts such as electricity-generation water, hardware manufacturing or training.
- Give a low/base/high range or IQR and identify inputs that could change it.
- Convert kWh and liters to monetary costs only with clearly identified local tariffs.
For example: “During May 2025, this text-inference workload used an estimated X Wh per request at the IT boundary, based on Y requests. Applying PUE Z gives A kWh at the facility boundary. Using a matching WUE of B L/kWh gives C liters of on-site cooling water. Electricity-generation water, hardware manufacturing and training are excluded. The range is D–E because utilization and output length were estimated.” Replace each letter with measured or explicitly assumed values.
What the published figures do—and do not—tell you
Google reports 0.03 gCO₂e for the median Gemini Apps text prompt using its 2024 average fleet grid carbon intensity. That figure is a carbon estimate, not a water or electricity price, and it uses the same provider-specific prompt population and accounting context described above.
Microsoft Datacenters reports a global PUE of 1.17 alongside its FY25 WUE of 0.27 L/kWh for the stated reporting population. Those fleet metrics can inform a scenario, but a different site, climate, cooling design or ownership boundary may produce another result.
Recommended Free Tools
No public source provides enough request-specific runtime, facility and lifecycle data to calculate an exact footprint for every commercial model request. The honest output is therefore a dated, scoped estimate with assumptions and unresolved inputs—not a single number presented as a law of AI.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
