Live estimate, not a meter reading
How much electricity and water AI data centers are using right now
There is no global meter for AI data centers. This counter uses IEA year anchors for all data centers, the OWID reading for 2025 AI use, derived AI values for 2024, and IEA water figures, then spreads them across the current second. The IEA itself treats today's totals as estimates.
Estimate since January 1 this year
Calendar year in UTC
Share of the UTC calendar year: 79.6%
Electricity
153.7 billion
kWh
6,123 kWh per second
Annual rate in this calendar year: 193.1 TWh
Water
53.8 billion
liters
2,142 liters per second
Annual rate in this calendar year: 67.5 billion L
Facilities with large numbers of accelerator chips for training and inference. That is the IEA’s definition. Water withdrawn for cooling that is not returned to its original source after use.
AI water is not measured separately. The counter derives it from the IEA water figure using AI’s share of data-center electricity. Hardware is excluded.
At a glance
The key figures and what they cover
Annual figures are estimates or projections. The live counter spreads them evenly over time and does not measure current facility use.
- All data centers, electricity in 2025
- 485 TWh
- IEA scenarios for all data centers in 2030 range from 833 TWh to 1,008 TWh. The table's base case is 945 TWh. The IEA executive summary rounds the same figure to around 950 TWh.
- AI-focused electricity in 2025
- 155 TWh
- OWID reading of an IEA figure. The IEA puts 2030 at around 465 TWh.
- Water use across all data centers in 2023
- 560 billion L
- The IEA projects this for 2030 1,200 billion L. The IEA total includes hardware, which the live counter excludes. The IEA does not publish an AI-only water total. The 2030 breakdown carries forward the 2023 mix.
For scale
What these volumes mean
- Households, using the IEA rule of thumb that 100 MW equals 100,000 homes
- 22,042,043
- Pools of 3.125 million liters in the year-to-date water total
- 17,201
- Counter electricity divided by German data-center electricity use in 2025 (Bitkom, 21.3 TWh)
- 9.1×
Jump to
Find the data you need
Go to the annual power and water trends, the scenarios and country data, or the optional calculations.
IEA curve
Electricity and water through 2030, year by year
The annual anchors come from the IEA reports. Between them sits the CAGR, not a meter reading. That is why 2026, and the live counter, already sit well above 2025.
Electricity for all data centers and AI-focused sites
The area between anchors is interpolated. All-DC 2024/2025/2030 and AI 2030 are IEA values. AI 2025 comes from the OWID reading, AI 2024 from +50 %. The dashed line marks the current calendar year.
2024
416 TWh
AI-focused: 103.3 TWh
2025
485 TWh
AI-focused: 155 TWh
2026
554.8 TWh
AI-focused: 193.1 TWh
2030
950 TWh
AI-focused: 465 TWh
IEA Key Questions on Energy and AI (2026)
From 416 TWh in 2024 to 950 TWh in 2030
AI share of all data-center electricity
The share is about 32% in 2025 and just under 49% in the IEA 2030 base case.
Electricity in TWh, water in billion liters
Two units, one timeline from 2024 to 2030. The water line reaches 1,200 billion liters in 2030.
gradually.aiWater in 2023 and 2030: volume and split
The IEA publishes the split only for 2023. 2030 uses that same mix, as the counter does. That is an assumption, not a new IEA figure.
2025 versus 2030
Germany 2025 against the world totals
Bitkom/Borderstep estimate German consumption, while the IEA estimates the world total. Their methods have not been fully reconciled.
gradually.aiProjections and regional data
How much confidence the consumption figures warrant
All data centers: four IEA scenarios for 2030
All four scenarios cover electricity use across data centers, including AI and other workloads. Table A.1 of the 2026 IEA report lists 945 TWh in the base case. The executive summary rounds this to around 950 TWh. These are two presentations of the same scenario. The scenarios describe possible pathways, not statistical confidence intervals or an AI-only range.
National data for 2025
Both series cover all data centers. The German study also includes smaller IT installations. The Irish statistics capture metered grid consumption. Their boundaries differ, so neither provides an AI-only total.
| Country | Data year | Electricity | Method and source | Share of national metered grid consumption |
|---|---|---|---|---|
| Germany | 2025 | 21.3 TWh | Modeled estimate | no comparable figure reported |
| Ireland | 2025 | 7.663 TWh | Metered grid consumption | 23% |
Lower water use per kWh does not guarantee a lower total
Microsoft's 2026 environmental report says its Phoenix data centers reduced water usage effectiveness by 23% in fiscal 2025. This describes those operations per unit of energy. It establishes neither a global water volume nor a decline in total AI data-center water consumption.
Microsoft, 2026 environmental report (fiscal 2025)Sources checked October 8, 2026
Show calculations and additional charts
Additional breakdowns
How the electricity and water figures break down
These views split the 2025 total into AI and the remainder, translate electricity into site equivalents, and distinguish water consumption from withdrawals. Values between annual anchors are calculated.
AI electricity from 2024 to 465 TWh
2024 follows from 155 TWh in 2025 and the IEA growth of 50%. The counter uses that same derivation.
AI and the rest, stacked over time
2025 is 155 TWh of AI load and 330 TWh without an AI focus. The CAGR sits between the anchors.
Electricity against water, year by year
Each point is a year. Water values from 2024 through 2029 sit on the CAGR between 560 and 1,200.
gradually.ai2025: 155 versus 330 TWh
The IEA total of 485 TWh splits into the OWID AI reading and the remainder.
Individual figures for text queries and agent tasks
The first three figures concern text queries but use different metrics and methods. The IEA agent examples count GPU electricity only. These figures do not support a ranking.
OpenAI, Epoch AI, Google Gemini Apps et al.
2023 water consumption by source
The published IEA split. This is not withdrawals.
Consumption versus withdrawals, 2023
A reading of IEA Figure 5.27 puts withdrawals at approximately 5,200 billion liters. This is an uncertain graph estimate. Published consumption is about 560 billion liters; only consumption enters the ticker.
Water through 2030, 2023 mix held constant
The IEA publishes the mix only for 2023. The areas after that scale the same shares. That is an assumption.
Total consumption and direct cooling
The direct line stays at 25% of the annual total because the mix is held fixed.
What the ticker takes from 2023
Direct 140, operational 513, including hardware 560. Hardware stays out of the ticker.
Milliliters of water per text query
Google estimates 0.26 ml of direct on-site consumptive cooling water for the median Gemini Apps prompt in May 2025. Sam Altman’s 0.000085-gallon figure has no published water boundary. These values are not methodologically harmonized and do not support a ranking.
Google Gemini Apps and OpenAI 2025
From withdrawals to direct cooling, 2023
One year, one source. Withdrawals are estimated from the graph; the other values are rounded consumption figures or derived shares. These are not AI-specific measurements.
IEA reference sites of 100 MW
Average annual power divided by 100 MW. That is an order of magnitude, not a site census.
gradually.aiGermany 2025 inside the world total
21.3 TWh is arithmetically about 4.4% of the 485 TWh worldwide. This comparison combines Bitkom/Borderstep and IEA estimates with different methods.
gradually.aiWhat the IEA means by one site
Rules of thumb from Energy and AI. The water illustration concerns an average US 100 MW data center and includes predominantly indirect water.
- Reference capacity
- 100 MW
- Household equivalent at 100 MW
- 100,000
- Water per day in the IEA illustration
- 2,000,000 L
- Households for that daily water volume
- 6,500
Arithmetic from the anchors
CAGR, per-second rates, and unit conversions
These tables recalculate annual anchors as growth rates, per-second averages, and query equivalents. AI-attributed water uses the same assumption shown in the counter.
Remainder without an AI focus, 2024 to 2030
Total minus AI. In 2025 that is the 330 TWh Our World in Data reads beside 155 TWh.
Water consumption from 560 to 1,200
Only the two IEA anchors. The CAGR between them is what the ticker uses.
Water in 2023 versus 2030
The IEA publishes the mix only for 2023. The 2030 direct line keeps that same 25%. That is an assumption.
2025 electricity and 2023 water, in one view
The 485 TWh total on the left. The published water split on the right.
Four years, five measures
All-DC 2024/2025/2030 and AI 2030 are IEA anchors. AI 2025 is the OWID reading, AI 2024 derived (+50 %). 2026 sits on the CAGR. Water in 2024 and 2026 sits between 560 and 1,200.
gradually.ai945 TWh versus 950 TWh
Our World in Data and Table A.1 of the 2026 IEA report list 945 TWh. The same report’s executive summary says around 950 TWh. The counter uses that rounded headline. This difference does not establish a revised forecast.
- OWID and IEA Table A.1, April 2026
- 945 TWh
- IEA executive summary, rounded, April 2026
- 950 TWh
- Gap
- 5 TWh
Withdrawals divided by consumption, 2023
The uncertain graph estimate of about 5,200 divided by about 560. The ratio is indicative; the ticker uses consumption only.
9.3×
IEA site water on a 365-day year
IEA illustration for an average US 100 MW data center. About 2 million liters consumed per day, more than 60% indirectly. Scaling to 365 days and comparing the global total gives a volume equivalent, not a site count.
- One site, liters in a 365-day year
- 730,000,000 L
- Such sites inside 560 billion liters
- 767.1
CAGR, per-second rates, and unit conversions · The tables show the same anchors as annual series, changes, and unit conversions.
Show tables (5)
Growth between the published years
A one-year span is a simple increase. Multi-year spans are CAGRs. The 50% AI rise in 2025 is an IEA figure.
| Span | Start | End | Years | Rate |
|---|---|---|---|---|
| All data centers, 2024-2025 | 416 TWh | 485 TWh | 1 | 16.6% |
| All data centers, 2025-2030 | 485 TWh | 950 TWh | 5 | 14.4% |
| AI-focused, 2024-2025 | 103.3 TWh | 155 TWh | 1 | 50% |
| AI-focused, 2025-2030 | 155 TWh | 465 TWh | 5 | 24.6% |
| Water consumption, 2023-2030 | 560 billion L | 1,200 billion L | 7 | 11.5% |
Electricity, every year from 2024 through 2030
All-DC 2024/2025/2030 and AI 2030 are IEA anchors. AI 2025 is the OWID reading, AI 2024 derived (+50 %). The years in between sit on the CAGR. Delta is the change from the previous year.
| Year | All data centers | AI-focused | Remainder without an AI focus | AI share | Delta, all, TWh | Delta, AI, TWh | kWh/s, all | kWh/s, AI | Times Germany 2025, all | Times Germany 2025, AI |
|---|---|---|---|---|---|---|---|---|---|---|
| 2024 | 416 TWh | 103.3 TWh | 312.7 TWh | 24.8% | - | - | 13,155.2 kWh/s | 3,267.7 kWh/s | 19.5× | 4.9× |
| 2025 | 485 TWh | 155 TWh | 330 TWh | 32% | 69 TWh | 51.7 TWh | 15,379.2 kWh/s | 4,915 kWh/s | 22.8× | 7.3× |
| 2026 | 554.8 TWh | 193.1 TWh | 361.7 TWh | 34.8% | 69.8 TWh | 38.1 TWh | 17,592.7 kWh/s | 6,122.8 kWh/s | 26× | 9.1× |
| 2027 | 634.7 TWh | 240.5 TWh | 394.1 TWh | 37.9% | 79.8 TWh | 47.4 TWh | 20,124.6 kWh/s | 7,627.3 kWh/s | 29.8× | 11.3× |
| 2028 | 726 TWh | 299.6 TWh | 426.3 TWh | 41.3% | 91.3 TWh | 59.1 TWh | 22,958.1 kWh/s | 9,475.7 kWh/s | 34.1× | 14.1× |
| 2029 | 830.5 TWh | 373.3 TWh | 457.2 TWh | 44.9% | 104.5 TWh | 73.6 TWh | 26,334.2 kWh/s | 11,836.5 kWh/s | 39× | 17.5× |
| 2030 | 950 TWh | 465 TWh | 485 TWh | 48.9% | 119.5 TWh | 91.7 TWh | 30,124.3 kWh/s | 14,745.1 kWh/s | 44.6× | 21.8× |
Water, every year from 2023 through 2030
The mix after 2023 is the 2023 split. AI columns start in 2024, when the electricity trajectory begins. AI 2024 is derived (+50 %), not an anchor. 2023 gets no AI share.
| Year | Total water consumption | Direct cooling | Energy supply | Hardware | Liters/s, total | AI direct, assumed | AI operational, assumed |
|---|---|---|---|---|---|---|---|
| 2023 | 560 billion L | 140 billion L | 373 billion L | 47 billion L | 17,757.5 L/s | no electricity anchor | no electricity anchor |
| 2024 | 624.4 billion L | 156.1 billion L | 415.9 billion L | 52.4 billion L | 19,745.9 L/s | 38.8 billion L | 142.1 billion L |
| 2025 | 696.2 billion L | 174.1 billion L | 463.7 billion L | 58.4 billion L | 22,077.6 L/s | 55.6 billion L | 203.8 billion L |
| 2026 | 776.3 billion L | 194.1 billion L | 517.1 billion L | 65.2 billion L | 24,617 L/s | 67.5 billion L | 247.5 billion L |
| 2027 | 865.6 billion L | 216.4 billion L | 576.6 billion L | 72.7 billion L | 27,448.6 L/s | 82 billion L | 300.5 billion L |
| 2028 | 965.2 billion L | 241.3 billion L | 642.9 billion L | 81 billion L | 30,522.3 L/s | 99.6 billion L | 364.9 billion L |
| 2029 | 1,076.2 billion L | 269.1 billion L | 716.8 billion L | 90.3 billion L | 34,126.4 L/s | 120.9 billion L | 443.1 billion L |
| 2030 | 1,200 billion L | 300 billion L | 799.3 billion L | 100.7 billion L | 38,051.8 L/s | 146.8 billion L | 538.1 billion L |
Queries in units of the published figures
1 kWh divided by published watt-hour references. This is a unit conversion, not a world-fleet load profile. Different methods do not support an efficiency ranking.
One UTC second of AI electricity as a calculated query equivalent
The 155 TWh AI electricity figure for 2025 is spread evenly across UTC seconds and divided by OpenAI’s 0.34 Wh mean. This is a unit conversion, not an observed query count or efficiency ranking.
14.5 million
Intensity and time
Multiples, liters per kWh, and hours
These views compare electricity, water, and time. Liters per kWh combines interpolated water totals with electricity figures from another IEA report; it is not a published IEA metric.
Where 2025 and 2023 sit against 2030
Electricity compares 2025 with 2030. Water compares 2023 with 2030, because that is where the water anchor sits.
gradually.aiAnchor end points, divided
465 TWh (IEA 2030) divided by 155 TWh (OWID reading 2025) is exactly 3. All-DC and water use IEA anchors.
gradually.aiRemainder without an AI focus, 2025 versus 2030
330 TWh in 2025 and 485 TWh in 2030. Both are total minus AI.
Liters per kWh, 2024 versus 2030
That year's water divided by that year's electricity. This is not an IEA intensity.
gradually.aiMultiples, liters per kWh, and hours · The radial and slope charts above show progress and multiples. Here are liters per kWh, rest growth, and time equivalents as tables.
Show tables (9)
Growth in the remainder only
2024 remainder is derived. 2025 and 2030 remainder sit in the OWID reading and the IEA difference.
| Span | Start | End | Years | Rate |
|---|---|---|---|---|
| Remainder without an AI focus, 2024-2025 | 312.7 TWh | 330 TWh | 1 | 5.5% |
| Remainder without an AI focus, 2025-2030 | 330 TWh | 485 TWh | 5 | 8% |
Intensity and year-on-year rates, 2024 through 2030
No 2023, because the electricity anchor starts later. YoY for all and AI follows each series CAGR from 2026. Rest YoY is the year-on-year growth of (all minus AI), not all YoY minus AI YoY, and is not a constant CAGR.
| Year | Liters/kWh, total | Liters/kWh, direct | YoY, all | YoY, AI | YoY, rest |
|---|---|---|---|---|---|
| 2024 | 1.5 L/kWh | 0.4 L/kWh | - | - | - |
| 2025 | 1.4 L/kWh | 0.4 L/kWh | 16.6% | 50% | 5.5% |
| 2026 | 1.4 L/kWh | 0.3 L/kWh | 14.4% | 24.6% | 9.6% |
| 2027 | 1.4 L/kWh | 0.3 L/kWh | 14.4% | 24.6% | 9% |
| 2028 | 1.3 L/kWh | 0.3 L/kWh | 14.4% | 24.6% | 8.2% |
| 2029 | 1.3 L/kWh | 0.3 L/kWh | 14.4% | 24.6% | 7.2% |
| 2030 | 1.3 L/kWh | 0.3 L/kWh | 14.4% | 24.6% | 6.1% |
Power, sites, and households, every year
Power and sites follow the 100 MW rule of thumb. The Germany column holds Bitkom 2025 fixed.
| Year | Average power, all | Average power, AI | 100 MW equivalents, all | 100 MW equivalents, AI | Household equivalents, all | Bitkom 2025 / world year |
|---|---|---|---|---|---|---|
| 2024 | 47,359 MW | 11,764 MW | 473.6 | 117.6 | 47,358,834 | 5.1% |
| 2025 | 55,365 MW | 17,694 MW | 553.7 | 176.9 | 55,365,297 | 4.4% |
| 2026 | 63,334 MW | 22,042 MW | 633.3 | 220.4 | 63,333,575 | 3.8% |
| 2027 | 72,449 MW | 27,458 MW | 724.5 | 274.6 | 72,448,664 | 3.4% |
| 2028 | 82,649 MW | 34,112 MW | 826.5 | 341.1 | 82,649,177 | 2.9% |
| 2029 | 94,803 MW | 42,611 MW | 948 | 426.1 | 94,803,228 | 2.6% |
| 2030 | 108,447 MW | 53,082 MW | 1,084.5 | 530.8 | 108,447,489 | 2.2% |
Water growth, sites, and pools
The rate after 2023 is the same CAGR every year. Site counts use the 365-day year. Pools per day use UTC days.
| Year | YoY, water | 365-day sites | Olympic pools | Pools per day |
|---|---|---|---|---|
| 2023 | - | 767.1 | 179,200 | 491 |
| 2024 | 11.5% | 855.4 | 199,813 | 545.9 |
| 2025 | 11.5% | 953.8 | 222,796 | 610.4 |
| 2026 | 11.5% | 1,063.5 | 248,423 | 680.6 |
| 2027 | 11.5% | 1,185.8 | 276,998 | 758.9 |
| 2028 | 11.5% | 1,322.2 | 308,860 | 843.9 |
| 2029 | 11.5% | 1,474.3 | 344,387 | 943.5 |
| 2030 | 11.5% | 1,643.8 | 384,000 | 1,052.1 |
Germany 2025 in hours of the world fleet
21.3 TWh from Bitkom 2025, placed on the 2025 world annual total. Two sources.
Query references relative to average AI electricity per second
The 155 TWh AI electricity figure for 2025 is spread evenly across UTC seconds. This unit conversion compares scale only, not observed request counts or an efficiency ranking.
One 100 MW site in a UTC year
100 MW across every second of the UTC year. 2024 and 2028 are leap years.
2023 withdrawals, rates only
2023 only, derived from an uncertain graph reading. All rates are indicative. No division by a later electricity total.
5 TWh as a share
The same gap, once against 945 and once against 950.
Scale and remainder
Daily averages, shares, and query equivalents
Daily averages and query equivalents recalculate annual figures. They show scale, not observed daily trends or load profiles.
485 TWh in 2025 as 100 dots
Each dot is 4.85 TWh. Colors match the AI/rest split used in the electricity charts. 155 is the OWID reading, 330 is the remainder (485 minus 155).
AI-to-rest ratio, 2024 vs 2030
2024: derived AI divided by rest. 2030: 465 AI divided by 485 rest (950 minus 465). A TWh ratio, not market share.
Daily averages, shares, and query equivalents · The tables add the distance to the 2030 anchors and the water breakdown to the daily averages.
Show tables (7)
Electricity as daily average, UTC year
Year TWh divided by UTC calendar days. No seasonality.
| Year | All, kWh/day | AI-focused, kWh/day | Rest, kWh/day |
|---|---|---|---|
| 2024 | 1,136,612,021.9 kWh/day | 282,331,511.8 kWh/day | 854,280,510 kWh/day |
| 2025 | 1,328,767,123.3 kWh/day | 424,657,534.2 kWh/day | 904,109,589 kWh/day |
| 2026 | 1,520,005,808.3 kWh/day | 529,009,029.2 kWh/day | 990,996,779.1 kWh/day |
| 2027 | 1,738,767,927.6 kWh/day | 659,002,915 kWh/day | 1,079,765,012.6 kWh/day |
| 2028 | 1,983,580,239.7 kWh/day | 818,697,314.1 kWh/day | 1,164,882,925.6 kWh/day |
| 2029 | 2,275,277,475.6 kWh/day | 1,022,670,756.8 kWh/day | 1,252,606,718.9 kWh/day |
| 2030 | 2,602,739,726 kWh/day | 1,273,972,602.7 kWh/day | 1,328,767,123.3 kWh/day |
Water as daily average, 2023 to 2030
Consumption in billion L, interpolated between 2023 and 2030.
| Year | Total, L/day | Direct, L/day |
|---|---|---|
| 2023 | 1,534,246,575.3 L/day | 383,561,643.8 L/day |
| 2024 | 1,706,049,733.3 L/day | 426,512,433.3 L/day |
| 2025 | 1,907,500,470.3 L/day | 476,875,117.6 L/day |
| 2026 | 2,126,911,401.2 L/day | 531,727,850.3 L/day |
| 2027 | 2,371,560,153.8 L/day | 592,890,038.5 L/day |
| 2028 | 2,637,124,718.3 L/day | 659,281,179.6 L/day |
| 2029 | 2,948,516,999.5 L/day | 737,129,249.9 L/day |
| 2030 | 3,287,671,232.9 L/day | 821,917,808.2 L/day |
AI share and ratio, every year
Share is AI divided by all. pp is the year-on-year change in percentage points.
| Year | AI share | Δ share | AI / rest |
|---|---|---|---|
| 2024 | 24.8% | - | 0.3× |
| 2025 | 32% | +7.1 pp | 0.5× |
| 2026 | 34.8% | +2.8 pp | 0.5× |
| 2027 | 37.9% | +3.1 pp | 0.6× |
| 2028 | 41.3% | +3.4 pp | 0.7× |
| 2029 | 44.9% | +3.7 pp | 0.8× |
| 2030 | 48.9% | +4 pp | 1× |
Calculated query equivalents for annual AI electricity
Each column divides the same AI annual total by a published watt-hour reference. These are not actual request counts and do not support an efficiency ranking because the methods differ.
| Year | ChatGPT, 0.34 Wh | Gemini, 0.24 Wh | Epoch, 0.3 Wh | Agent, 1.14 Wh | Reasoning, 50 Wh |
|---|---|---|---|---|---|
| 2024 | 303,921.6 billion | 430,555.6 billion | 344,444.4 billion | 90,643.3 billion | 2,066.7 billion |
| 2025 | 455,882.4 billion | 645,833.3 billion | 516,666.7 billion | 135,964.9 billion | 3,100 billion |
| 2026 | 567,906.8 billion | 804,534.6 billion | 643,627.7 billion | 169,375.7 billion | 3,861.8 billion |
| 2027 | 707,459 billion | 1,002,233.6 billion | 801,786.9 billion | 210,996.5 billion | 4,810.7 billion |
| 2028 | 881,303.6 billion | 1,248,513.4 billion | 998,810.7 billion | 262,844.9 billion | 5,992.9 billion |
| 2029 | 1,097,867.1 billion | 1,555,311.8 billion | 1,244,249.4 billion | 327,434.1 billion | 7,465.5 billion |
| 2030 | 1,367,647.1 billion | 1,937,500 billion | 1,550,000 billion | 407,894.7 billion | 9,300 billion |
Still missing by 2030
2030 anchor minus current year. Water only from 2023 where the anchor starts.
| Year | All, TWh missing | AI, TWh missing | Rest, TWh missing | Water, bn L missing |
|---|---|---|---|---|
| 2024 | 534 TWh | 361.7 TWh | 172.3 TWh | 575.6 billion L |
| 2025 | 465 TWh | 310 TWh | 155 TWh | 503.8 billion L |
| 2026 | 395.2 TWh | 271.9 TWh | 123.3 TWh | 423.7 billion L |
| 2027 | 315.3 TWh | 224.5 TWh | 90.9 TWh | 334.4 billion L |
| 2028 | 224 TWh | 165.4 TWh | 58.7 TWh | 234.8 billion L |
| 2029 | 119.5 TWh | 91.7 TWh | 27.8 TWh | 123.8 billion L |
| 2030 | 0 TWh | 0 TWh | 0 TWh | 0 billion L |
Anchor snapshot 2024, 2025, 2030
One row per metric. Water in 2024 is interpolated, not a standalone anchor.
| Metric | 2024 | 2025 | 2030 |
|---|---|---|---|
| All data centers, TWh | 416 TWh | 485 TWh | 950 TWh |
| AI-focused, TWh | 103.3 TWh | 155 TWh | 465 TWh |
| Rest without AI, TWh | 312.7 TWh | 330 TWh | 485 TWh |
| AI share | 24.8% | 32% | 48.9% |
| Water, bn L | 624.4 billion L | 696.2 billion L | 1,200 billion L |
| Power, MW | 47,359 MW | 55,365 MW | 108,447 MW |
Water mix shares, fixed
140 / 373 / 47 from the IEA water report. Every row is identical because the mix is fixed.
| Year | Direct | Power/heat | Hardware |
|---|---|---|---|
| 2023 | 25% | 66.6% | 8.4% |
| 2024 | 25% | 66.6% | 8.4% |
| 2025 | 25% | 66.6% | 8.4% |
| 2026 | 25% | 66.6% | 8.4% |
| 2027 | 25% | 66.6% | 8.4% |
| 2028 | 25% | 66.6% | 8.4% |
| 2029 | 25% | 66.6% | 8.4% |
| 2030 | 25% | 66.6% | 8.4% |
Index and pace
Progress and uniform monthly averages
Monthly and weekly figures divide annual anchors by the number of months or weeks. They are calculated averages, not time-resolved measurements or demand profiles.
485 TWh in 2025, spread evenly across 12 months
Each month = annual total divided by 12. No season, just another grid.
155 TWh AI in 2025, spread evenly across 12 months
Same assumption as above, for the AI-focused share only.
gradually.ai2030 vs index base (power 2024, water 2023)
100 = base year. Water in 2030 is more than double 2023.
gradually.aiShare of 2030 anchors already reached
Power compares 2024 to 2030. Water compares 2023 to 2030. No new anchors.
Water progress, 2023 vs 2030
Power compares 2024 to 2030. Water compares 2023 to 2030. No new anchors.
Progress and uniform monthly averages · The tables express annual anchors as monthly, weekly, and progress values. They do not add new measurements to the scenarios.
Show tables (7)
Index 2024 = 100 (power), 2023 = 100 (water)
2023 only has the water index. No electricity anchor there.
| Year | All, index | AI, index | Rest, index | Water, index |
|---|---|---|---|---|
| 2023 | - | - | - | 100 index |
| 2024 | 100 index | 100 index | 100 index | 111.5 index |
| 2025 | 116.6 index | 150 index | 105.5 index | 124.3 index |
| 2026 | 133.4 index | 186.9 index | 115.7 index | 138.6 index |
| 2027 | 152.6 index | 232.8 index | 126 index | 154.6 index |
| 2028 | 174.5 index | 290 index | 136.4 index | 172.4 index |
| 2029 | 199.6 index | 361.2 index | 146.2 index | 192.2 index |
| 2030 | 228.4 index | 450 index | 155.1 index | 214.3 index |
Electricity as monthly average, uniform
TWh divided by 12. No calendar month profile.
| Year | All, TWh/month | AI, TWh/month | Rest, TWh/month |
|---|---|---|---|
| 2024 | 34.7 TWh/month | 8.6 TWh/month | 26.1 TWh/month |
| 2025 | 40.4 TWh/month | 12.9 TWh/month | 27.5 TWh/month |
| 2026 | 46.2 TWh/month | 16.1 TWh/month | 30.1 TWh/month |
| 2027 | 52.9 TWh/month | 20 TWh/month | 32.8 TWh/month |
| 2028 | 60.5 TWh/month | 25 TWh/month | 35.5 TWh/month |
| 2029 | 69.2 TWh/month | 31.1 TWh/month | 38.1 TWh/month |
| 2030 | 79.2 TWh/month | 38.8 TWh/month | 40.4 TWh/month |
Water as monthly average, 2023 to 2030
Billion liters divided by 12 at a fixed mix.
| Year | Total, bn L/month | Direct, bn L/month | Power/heat, bn L/month | Hardware, bn L/month |
|---|---|---|---|---|
| 2023 | 46.7 bn L/month | 11.7 bn L/month | 31.1 bn L/month | 3.9 bn L/month |
| 2024 | 52 bn L/month | 13 bn L/month | 34.7 bn L/month | 4.4 bn L/month |
| 2025 | 58 bn L/month | 14.5 bn L/month | 38.6 bn L/month | 4.9 bn L/month |
| 2026 | 64.7 bn L/month | 16.2 bn L/month | 43.1 bn L/month | 5.4 bn L/month |
| 2027 | 72.1 bn L/month | 18 bn L/month | 48 bn L/month | 6.1 bn L/month |
| 2028 | 80.4 bn L/month | 20.1 bn L/month | 53.6 bn L/month | 6.8 bn L/month |
| 2029 | 89.7 bn L/month | 22.4 bn L/month | 59.7 bn L/month | 7.5 bn L/month |
| 2030 | 100 bn L/month | 25 bn L/month | 66.6 bn L/month | 8.4 bn L/month |
Electricity as weekly average, UTC year
TWh divided by UTC days/7. Leap years change the week count slightly, not the 7-day week.
| Year | All, TWh/week | AI, TWh/week | Rest, TWh/week |
|---|---|---|---|
| 2024 | 8 TWh/week | 2 TWh/week | 6 TWh/week |
| 2025 | 9.3 TWh/week | 3 TWh/week | 6.3 TWh/week |
| 2026 | 10.6 TWh/week | 3.7 TWh/week | 6.9 TWh/week |
| 2027 | 12.2 TWh/week | 4.6 TWh/week | 7.6 TWh/week |
| 2028 | 13.9 TWh/week | 5.7 TWh/week | 8.2 TWh/week |
| 2029 | 15.9 TWh/week | 7.2 TWh/week | 8.8 TWh/week |
| 2030 | 18.2 TWh/week | 8.9 TWh/week | 9.3 TWh/week |
Progress to 2030, every year
Percent = year value divided by 2030 anchor. Water from 2023.
| Year | All, % of 2030 | AI, % of 2030 | Water, % of 2030 |
|---|---|---|---|
| 2023 | - | - | 46.7% |
| 2024 | 43.8% | 22.2% | 52% |
| 2025 | 51.1% | 33.3% | 58% |
| 2026 | 58.4% | 41.5% | 64.7% |
| 2027 | 66.8% | 51.7% | 72.1% |
| 2028 | 76.4% | 64.4% | 80.4% |
| 2029 | 87.4% | 80.3% | 89.7% |
| 2030 | 100% | 100% | 100% |
Rest power and 100 MW sites
Rest MW = all MW minus AI MW. Sites follow the 100 MW rule of thumb.
| Year | Rest, MW | Rest, 100 MW sites |
|---|---|---|
| 2024 | 35,595 MW | 356 |
| 2025 | 37,671 MW | 376.7 |
| 2026 | 41,292 MW | 412.9 |
| 2027 | 44,990 MW | 449.9 |
| 2028 | 48,537 MW | 485.4 |
| 2029 | 52,192 MW | 521.9 |
| 2030 | 55,365 MW | 553.7 |
AI-attributed water, 2024 to 2030
AI share of year water times mix. 2023 omitted because no AI power anchor.
| Year | AI direct, bn L | AI operational, bn L |
|---|---|---|
| 2024 | 38.8 billion L | 142.1 billion L |
| 2025 | 55.6 billion L | 203.8 billion L |
| 2026 | 67.5 billion L | 247.5 billion L |
| 2027 | 82 billion L | 300.5 billion L |
| 2028 | 99.6 billion L | 364.9 billion L |
| 2029 | 120.9 billion L | 443.1 billion L |
| 2030 | 146.8 billion L | 538.1 billion L |
Benchmarks and ratios
Ratios, hours, Germany, and remainder to 2030
These metrics compare annual totals. The only German figure here is Bitkom's 21.3 TWh estimate for 2025, not a German time series.
Electricity ratios, 2024 vs 2030
All/AI and rest/AI as pure TWh quotients. Not market share.
World year divided by Germany 2025
21.3 TWh Bitkom 2025 in the denominator, each world year in the numerator. Germany does not grow here.
gradually.aiRatios, hours, Germany, and remainder to 2030 · The tables show hourly averages, UTC time bases, and calculated query equivalents.
Show tables (10)
Ratios per calendar year
All/AI, rest/AI, and all/rest from published, derived, and CAGR-interpolated TWh.
| Year | All / AI | Rest / AI | All / rest |
|---|---|---|---|
| 2024 | 4× | 3× | 1.3× |
| 2025 | 3.1× | 2.1× | 1.5× |
| 2026 | 2.9× | 1.9× | 1.5× |
| 2027 | 2.6× | 1.6× | 1.6× |
| 2028 | 2.4× | 1.4× | 1.7× |
| 2029 | 2.2× | 1.2× | 1.8× |
| 2030 | 2× | 1× | 2× |
Electricity as hourly average, UTC year
TWh divided by UTC hours. No intraday profile.
| Year | All, kWh/h | AI, kWh/h | Rest, kWh/h |
|---|---|---|---|
| 2024 | 47,358,834.2 kWh/h | 11,763,813 kWh/h | 35,595,021.3 kWh/h |
| 2025 | 55,365,296.8 kWh/h | 17,694,063.9 kWh/h | 37,671,232.9 kWh/h |
| 2026 | 63,333,575.3 kWh/h | 22,042,042.9 kWh/h | 41,291,532.5 kWh/h |
| 2027 | 72,448,663.6 kWh/h | 27,458,454.8 kWh/h | 44,990,208.9 kWh/h |
| 2028 | 82,649,176.7 kWh/h | 34,112,388.1 kWh/h | 48,536,788.6 kWh/h |
| 2029 | 94,803,228.2 kWh/h | 42,611,281.5 kWh/h | 52,191,946.6 kWh/h |
| 2030 | 108,447,488.6 kWh/h | 53,082,191.8 kWh/h | 55,365,296.8 kWh/h |
Water as hourly average, 2023 to 2030
Billion liters divided by UTC hours.
| Year | Total, L/h |
|---|---|
| 2023 | 63,926,940.6 L/h |
| 2024 | 71,085,405.6 L/h |
| 2025 | 79,479,186.3 L/h |
| 2026 | 88,621,308.4 L/h |
| 2027 | 98,815,006.4 L/h |
| 2028 | 109,880,196.6 L/h |
| 2029 | 122,854,875 L/h |
| 2030 | 136,986,301.4 L/h |
Absolute year-on-year gains
TWh or billion liters vs prior year. Power deltas start in 2025 because the anchor begins in 2024.
| Year | All, Δ TWh | AI, Δ TWh | Rest, Δ TWh | Water, Δ bn L |
|---|---|---|---|---|
| 2023 | - | - | - | - |
| 2024 | - | - | - | 64.4 billion L |
| 2025 | 69 TWh | 51.7 TWh | 17.3 TWh | 71.8 billion L |
| 2026 | 69.8 TWh | 38.1 TWh | 31.7 TWh | 80.1 billion L |
| 2027 | 79.8 TWh | 47.4 TWh | 32.4 TWh | 89.3 billion L |
| 2028 | 91.3 TWh | 59.1 TWh | 32.2 TWh | 99.6 billion L |
| 2029 | 104.5 TWh | 73.6 TWh | 30.9 TWh | 111 billion L |
| 2030 | 119.5 TWh | 91.7 TWh | 27.8 TWh | 123.8 billion L |
Germany 2025 as fixed divisor
World/AI/rest divided by 21.3 TWh. Only the denominator is country-specific.
| Year | World / DE | AI / DE | Rest / DE |
|---|---|---|---|
| 2024 | 19.5× | 4.9× | 14.7× |
| 2025 | 22.8× | 7.3× | 15.5× |
| 2026 | 26× | 9.1× | 17× |
| 2027 | 29.8× | 11.3× | 18.5× |
| 2028 | 34.1× | 14.1× | 20× |
| 2029 | 39× | 17.5× | 21.5× |
| 2030 | 44.6× | 21.8× | 22.8× |
Olympic pools and pools per day
Reference pool of 3.125 million liters (50 × 25 × 2.5 m, minimum dimensions of a permanent Olympic swimming pool, World Aquatics 2026). Growth follows the water CAGR.
| Year | Pools total | Pools / day |
|---|---|---|
| 2023 | 179,200 | 491 |
| 2024 | 199,813 | 545.9 |
| 2025 | 222,796 | 610.4 |
| 2026 | 248,423 | 680.6 |
| 2027 | 276,998 | 758.9 |
| 2028 | 308,860 | 843.9 |
| 2029 | 344,387 | 943.5 |
| 2030 | 384,000 | 1,052.1 |
Calculated query equivalents per AI TWh, 2024 to 2030
Scaled against the AI electricity annual anchor. A hypothetical unit conversion, not actual request counts or a load profile. Different methods do not support an efficiency ranking.
| Year | ChatGPT | Gemini | Epoch | Agent | Reasoning |
|---|---|---|---|---|---|
| 2024 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2025 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2026 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2027 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2028 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2029 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
| 2030 | 2,941.2 billion | 4,166.7 billion | 3,333.3 billion | 877.2 billion | 20 billion |
Remainder to 2030 in percent
100 % minus the share of the 2030 anchor already reached. Water from 2023.
| Year | All, % missing | AI, % missing | Water, % missing |
|---|---|---|---|
| 2023 | - | - | 53.3% |
| 2024 | 56.2% | 77.8% | 48% |
| 2025 | 48.9% | 66.7% | 42% |
| 2026 | 41.6% | 58.5% | 35.3% |
| 2027 | 33.2% | 48.3% | 27.9% |
| 2028 | 23.6% | 35.6% | 19.6% |
| 2029 | 12.6% | 19.7% | 10.3% |
| 2030 | 0% | 0% | 0% |
UTC calendar per year
Transparency for days, hours, and weeks behind every rate.
| Year | Days | Hours | Seconds | Weeks |
|---|---|---|---|---|
| 2023 | 365 | 8,760 | 31,536,000 | 52.1 |
| 2024 | 366 | 8,784 | 31,622,400 | 52.3 |
| 2025 | 365 | 8,760 | 31,536,000 | 52.1 |
| 2026 | 365 | 8,760 | 31,536,000 | 52.1 |
| 2027 | 365 | 8,760 | 31,536,000 | 52.1 |
| 2028 | 366 | 8,784 | 31,622,400 | 52.3 |
| 2029 | 365 | 8,760 | 31,536,000 | 52.1 |
| 2030 | 365 | 8,760 | 31,536,000 | 52.1 |
100 MW-years and site equivalents
One 100 MW-year split into all data centers, AI-focused, and the remainder without an AI focus.
| Year | 100 MW-year, TWh | All, 100 MW-years | AI, 100 MW-years | Rest, 100 MW-years | All, 100 MW sites |
|---|---|---|---|---|---|
| 2024 | 0.9 TWh | 473.6 | 117.6 | 356 | 473.6 |
| 2025 | 0.9 TWh | 553.7 | 176.9 | 376.7 | 553.7 |
| 2026 | 0.9 TWh | 633.3 | 220.4 | 412.9 | 633.3 |
| 2027 | 0.9 TWh | 724.5 | 274.6 | 449.9 | 724.5 |
| 2028 | 0.9 TWh | 826.5 | 341.1 | 485.4 | 826.5 |
| 2029 | 0.9 TWh | 948 | 426.1 | 521.9 | 948 |
| 2030 | 0.9 TWh | 1,084.5 | 530.8 | 553.7 | 1,084.5 |
Composition and micro-time
Growth shares, minutes, seconds, and AI water
These views break down the change between annual anchors. Minute and second figures are uniform averages, not intraday profiles.
Growth 2024 to 2030: AI plus rest
416 TWh start, plus AI and rest delta to 950 TWh. Pure TWh bridge.
AI share and AI/rest ratio
Bar: AI share of total power in %. Line: AI divided by rest, scaled to 100.
Seconds per TWh in 2025
UTC year seconds divided by 2025 TWh (485 all-DC IEA, 155 AI OWID reading, rest derived). Uniform annual pace.
Growth shares, minutes, seconds, and AI water · The tables place annual growth contributions and time averages alongside the charts.
Show tables (8)
Electricity as minute average, UTC year
TWh divided by UTC minutes. No intraday profile.
| Year | All, kWh/min | AI, kWh/min | Rest, kWh/min |
|---|---|---|---|
| 2024 | 789,313.9 kWh/min | 196,063.5 kWh/min | 593,250.4 kWh/min |
| 2025 | 922,754.9 kWh/min | 294,901.1 kWh/min | 627,853.9 kWh/min |
| 2026 | 1,055,559.6 kWh/min | 367,367.4 kWh/min | 688,192.2 kWh/min |
| 2027 | 1,207,477.7 kWh/min | 457,640.9 kWh/min | 749,836.8 kWh/min |
| 2028 | 1,377,486.3 kWh/min | 568,539.8 kWh/min | 808,946.5 kWh/min |
| 2029 | 1,580,053.8 kWh/min | 710,188 kWh/min | 869,865.8 kWh/min |
| 2030 | 1,807,458.1 kWh/min | 884,703.2 kWh/min | 922,754.9 kWh/min |
Water as minute average, 2023 to 2030
Billion liters divided by UTC minutes.
| Year | Total, L/min | Direct, L/min |
|---|---|---|
| 2023 | 1,065,449 L/min | 266,362.3 L/min |
| 2024 | 1,184,756.8 L/min | 296,189.2 L/min |
| 2025 | 1,324,653.1 L/min | 331,163.3 L/min |
| 2026 | 1,477,021.8 L/min | 369,255.5 L/min |
| 2027 | 1,646,916.8 L/min | 411,729.2 L/min |
| 2028 | 1,831,336.6 L/min | 457,834.2 L/min |
| 2029 | 2,047,581.2 L/min | 511,895.3 L/min |
| 2030 | 2,283,105 L/min | 570,776.3 L/min |
Seconds per TWh, 2024 to 2030
How long one TWh takes at an even annual pace.
| Year | All | AI | Rest |
|---|---|---|---|
| 2024 | 76,015.4 s/TWh | 306,023.2 s/TWh | 101,137.7 s/TWh |
| 2025 | 65,022.7 s/TWh | 203,458.1 s/TWh | 95,563.6 s/TWh |
| 2026 | 56,841.9 s/TWh | 163,324.2 s/TWh | 87,184.9 s/TWh |
| 2027 | 49,690.4 s/TWh | 131,107.2 s/TWh | 80,017.4 s/TWh |
| 2028 | 43,557.6 s/TWh | 105,533.5 s/TWh | 74,170.5 s/TWh |
| 2029 | 37,973.4 s/TWh | 84,484.7 s/TWh | 68,976.2 s/TWh |
| 2030 | 33,195.8 s/TWh | 67,819.4 s/TWh | 65,022.7 s/TWh |
Who delivers the year-on-year growth?
Δ TWh vs prior year. AI share = AI Δ divided by total Δ.
| Year | AI, Δ TWh | Rest, Δ TWh | All, Δ TWh | AI share of Δ |
|---|---|---|---|---|
| 2024 | - | - | - | - |
| 2025 | 51.7 TWh | 17.3 TWh | 69 TWh | 74.9% |
| 2026 | 38.1 TWh | 31.7 TWh | 69.8 TWh | 54.6% |
| 2027 | 47.4 TWh | 32.4 TWh | 79.8 TWh | 59.4% |
| 2028 | 59.1 TWh | 32.2 TWh | 91.3 TWh | 64.7% |
| 2029 | 73.6 TWh | 30.9 TWh | 104.5 TWh | 70.5% |
| 2030 | 91.7 TWh | 27.8 TWh | 119.5 TWh | 76.7% |
Total span 2024 to 2030
Sum of year deltas as a check on the waterfall bridge.
| All, Δ TWh total | AI, Δ TWh total | Rest, Δ TWh total | AI share of total Δ | Rest share of total Δ |
|---|---|---|---|---|
| 534 TWh | 361.7 TWh | 172.3 TWh | 67.7% | 32.3% |
Germany 2025 as a share of world
21.3 TWh divided by that year's world/AI/rest TWh. No DE growth.
| Year | DE / world, % | DE / AI, % | DE / rest, % |
|---|---|---|---|
| 2024 | 5.1% | 20.6% | 6.8% |
| 2025 | 4.4% | 13.7% | 6.5% |
| 2026 | 3.8% | 11% | 5.9% |
| 2027 | 3.4% | 8.9% | 5.4% |
| 2028 | 2.9% | 7.1% | 5% |
| 2029 | 2.6% | 5.7% | 4.7% |
| 2030 | 2.2% | 4.6% | 4.4% |
AI-attributed water per AI TWh
2023 water mix scaled by AI share. No 2023 AI water as fact.
| Year | Direct, bn L AI/year | Operational, bn L AI/year | Direct, L/kWh AI | Operational, L/kWh AI |
|---|---|---|---|---|
| 2024 | 38.8 billion L | 142.1 billion L | 0.4 L/kWh | 1.4 L/kWh |
| 2025 | 55.6 billion L | 203.8 billion L | 0.4 L/kWh | 1.3 L/kWh |
| 2026 | 67.5 billion L | 247.5 billion L | 0.3 L/kWh | 1.3 L/kWh |
| 2027 | 82 billion L | 300.5 billion L | 0.3 L/kWh | 1.2 L/kWh |
| 2028 | 99.6 billion L | 364.9 billion L | 0.3 L/kWh | 1.2 L/kWh |
| 2029 | 120.9 billion L | 443.1 billion L | 0.3 L/kWh | 1.2 L/kWh |
| 2030 | 146.8 billion L | 538.1 billion L | 0.3 L/kWh | 1.2 L/kWh |
Reference values as a share of average AI electricity per UTC second in 2025
Query watt-hours divided by global AI electricity per UTC second at an even pace. This compares scale, not simultaneous use or relative efficiency.
| ChatGPT | Gemini | Epoch | Agent | Reasoning |
|---|---|---|---|---|
| 0.00000692% | 0.00000488% | 0.0000061% | 0.00002319% | 0.00101729% |
Cadence and coupling
Quarters, households, and a withdrawal proxy
The tables recalculate annual figures as quarter and week averages. Withdrawals appear only as a 2023 ratio applied to consumption, not as a time series.
Household equivalents, 2024 vs 2030
IEA: 100 MW AI DC = 100,000 households. Scales with interpolated MW.
MW per TWh and liters per kWh
Left axis: average MW divided by TWh. Right axis: water in bn L divided by power TWh (i.e. L/kWh).
gradually.aiStack shares in 2030
AI and rest share of total power in percent, IEA 2030 anchor.
Quarters, households, and a withdrawal proxy · The tables show quarter and week averages, household equivalents, and the share without an AI focus.
Show tables (9)
Electricity as quarterly average, UTC year
Annual TWh divided by 4. No quarterly profile.
| Year | All, TWh/quarter | AI, TWh/quarter | Rest, TWh/quarter |
|---|---|---|---|
| 2024 | 104 TWh/quarter | 25.8 TWh/quarter | 78.2 TWh/quarter |
| 2025 | 121.3 TWh/quarter | 38.8 TWh/quarter | 82.5 TWh/quarter |
| 2026 | 138.7 TWh/quarter | 48.3 TWh/quarter | 90.4 TWh/quarter |
| 2027 | 158.7 TWh/quarter | 60.1 TWh/quarter | 98.5 TWh/quarter |
| 2028 | 181.5 TWh/quarter | 74.9 TWh/quarter | 106.6 TWh/quarter |
| 2029 | 207.6 TWh/quarter | 93.3 TWh/quarter | 114.3 TWh/quarter |
| 2030 | 237.5 TWh/quarter | 116.3 TWh/quarter | 121.3 TWh/quarter |
Water as quarterly average, 2023 to 2030
Annual consumption divided by 4. From 2024, the 2023 water mix is held fixed. This is not a quarterly profile.
| Year | Total, bn L/quarter |
|---|---|
| 2023 | 140 bn L/quarter |
| 2024 | 156.1 bn L/quarter |
| 2025 | 174.1 bn L/quarter |
| 2026 | 194.1 bn L/quarter |
| 2027 | 216.4 bn L/quarter |
| 2028 | 241.3 bn L/quarter |
| 2029 | 269.1 bn L/quarter |
| 2030 | 300 bn L/quarter |
Water as weekly average, UTC year
Annual consumption divided by UTC weeks. This is not a weekly profile.
| Year | Total, bn L/week |
|---|---|
| 2023 | 10.7 bn L/week |
| 2024 | 11.9 bn L/week |
| 2025 | 13.4 bn L/week |
| 2026 | 14.9 bn L/week |
| 2027 | 16.6 bn L/week |
| 2028 | 18.5 bn L/week |
| 2029 | 20.6 bn L/week |
| 2030 | 23 bn L/week |
Household equivalents per calendar year
All and AI from MW path and IEA reference.
| Year | All, households | AI, households |
|---|---|---|
| 2024 | 47,358,834 | 11,763,813 |
| 2025 | 55,365,297 | 17,694,064 |
| 2026 | 63,333,575 | 22,042,043 |
| 2027 | 72,448,664 | 27,458,455 |
| 2028 | 82,649,177 | 34,112,388 |
| 2029 | 94,803,228 | 42,611,282 |
| 2030 | 108,447,489 | 53,082,192 |
TWh per UTC second
Year TWh (anchor, derived, or CAGR) divided by UTC year seconds.
| Year | All, TWh/s | AI, TWh/s | Rest, TWh/s |
|---|---|---|---|
| 2024 | 0.00001316 TWh/s | 0.000003268 TWh/s | 0.000009888 TWh/s |
| 2025 | 0.00001538 TWh/s | 0.000004915 TWh/s | 0.00001046 TWh/s |
| 2026 | 0.00001759 TWh/s | 0.000006123 TWh/s | 0.00001147 TWh/s |
| 2027 | 0.00002012 TWh/s | 0.000007627 TWh/s | 0.0000125 TWh/s |
| 2028 | 0.00002296 TWh/s | 0.000009476 TWh/s | 0.00001348 TWh/s |
| 2029 | 0.00002633 TWh/s | 0.00001184 TWh/s | 0.0000145 TWh/s |
| 2030 | 0.00003012 TWh/s | 0.00001475 TWh/s | 0.00001538 TWh/s |
Rest share of total power
100 % minus AI share, pure TWh quote.
| Year | AI, % | Rest, % |
|---|---|---|
| 2024 | 24.8% | 75.2% |
| 2025 | 32% | 68% |
| 2026 | 34.8% | 65.2% |
| 2027 | 37.9% | 62.1% |
| 2028 | 41.3% | 58.7% |
| 2029 | 44.9% | 55.1% |
| 2030 | 48.9% | 51.1% |
YoY growth AI vs rest
Ratio of year-on-year growth rates. Null when rest does not grow.
| Year | AI, YoY | Rest, YoY | AI/rest YoY |
|---|---|---|---|
| 2024 | - | - | - |
| 2025 | 50% | 5.5% | 9× |
| 2026 | 24.6% | 9.6% | 2.6× |
| 2027 | 24.6% | 9% | 2.7× |
| 2028 | 24.6% | 8.2% | 3× |
| 2029 | 24.6% | 7.2% | 3.4× |
| 2030 | 24.6% | 6.1% | 4× |
Water Δ to power Δ
Billion liters Δ divided by TWh Δ vs prior year. From 2025 only.
| Year | Water, Δ bn L | All, Δ TWh | Δwater/Δpower |
|---|---|---|---|
| 2024 | - | - | - |
| 2025 | 71.8 billion L | 69 TWh | 1 |
| 2026 | 80.1 billion L | 69.8 TWh | 1.1 |
| 2027 | 89.3 billion L | 79.8 TWh | 1.1 |
| 2028 | 99.6 billion L | 91.3 TWh | 1.1 |
| 2029 | 111 billion L | 104.5 TWh | 1.1 |
| 2030 | 123.8 billion L | 119.5 TWh | 1 |
Withdrawal proxy from 2023 ratio
Uncertain graph reading 5,200/560 applied as a scalar to consumption CAGR. Every proxy inherits the reading uncertainty; this is not a measured withdrawal series.
| Year | Consumption, bn L | Withdrawal proxy, bn L | 2023 ratio |
|---|---|---|---|
| 2023 | 560 billion L | 5,200 billion L | 9.3× |
| 2024 | 624.4 billion L | 5,798.1 billion L | 9.3× |
| 2025 | 696.2 billion L | 6,465.1 billion L | 9.3× |
| 2026 | 776.3 billion L | 7,208.7 billion L | 9.3× |
| 2027 | 865.6 billion L | 8,037.9 billion L | 9.3× |
| 2028 | 965.2 billion L | 8,962.5 billion L | 9.3× |
| 2029 | 1,076.2 billion L | 9,993.4 billion L | 9.3× |
| 2030 | 1,200 billion L | 11,142.9 billion L | 9.3× |
Anchor ledger and references
Anchor and derived constants, half-years, and site benchmarks
This section lists published anchors and the constants behind the calculations. The additional charts do not show further measurements.
MW stack, 2024 vs 2030
AI MW and rest MW from UTC annual pace. Rest = all minus AI.
2025 TWh comparison (symmetric)
AI, rest, and Bitkom 21.3 TWh as nodes. Ribbon width = minimum of the two TWh.
gradually.ai2023 water mix in billion liters
140/373/47 from IEA Energy and AI. No CAGR on mix shares.
950 TWh in 2030, spread evenly per month
950 TWh IEA 2030 anchor, divided by 12. No seasonality.
465 TWh AI in 2030, spread evenly per month
465 TWh IEA 2030 anchor, divided by 12.
gradually.aiAnchor and derived constants, half-years, and site benchmarks · The tables bring together half-year values, power shares, and other constants from the same annual anchors.
Show tables (10)
Anchor values and derived constants
IEA, OWID, Bitkom, and query cards. AI 2025 is the OWID reading of IEA Fig. 1.5. AI 2024 is back-calculated from +50%, not a typed IEA numeral.
| Metric | Value |
|---|---|
| All 2024, TWh | 416 TWh |
| All 2025, TWh | 485 TWh |
| All 2030, TWh | 950 TWh |
| AI 2024, TWh | 103.3 TWh |
| AI 2025, TWh | 155 TWh |
| AI 2030, TWh | 465 TWh |
| Water 2023, bn L | 560 billion L |
| Water 2030, bn L | 1,200 billion L |
| OWID 2030, TWh | 945 TWh |
| IEA 2030, TWh | 950 TWh |
| DE 2025, TWh | 21.3 TWh |
| ChatGPT, Wh | 0.3 |
| Gemini, ml | 0.3 |
Electricity as weekly average, UTC year
Annual figures divided by the weeks in the UTC calendar year.
| Year | All, TWh/week | AI, TWh/week | Rest, TWh/week |
|---|---|---|---|
| 2024 | 8 TWh/week | 2 TWh/week | 6 TWh/week |
| 2025 | 9.3 TWh/week | 3 TWh/week | 6.3 TWh/week |
| 2026 | 10.6 TWh/week | 3.7 TWh/week | 6.9 TWh/week |
| 2027 | 12.2 TWh/week | 4.6 TWh/week | 7.6 TWh/week |
| 2028 | 13.9 TWh/week | 5.7 TWh/week | 8.2 TWh/week |
| 2029 | 15.9 TWh/week | 7.2 TWh/week | 8.8 TWh/week |
| 2030 | 18.2 TWh/week | 8.9 TWh/week | 9.3 TWh/week |
Electricity as half-year average
Annual TWh divided by 2.
| Year | All, TWh/half-year | AI, TWh/half-year | Rest, TWh/half-year |
|---|---|---|---|
| 2024 | 208 TWh/half-year | 51.7 TWh/half-year | 156.3 TWh/half-year |
| 2025 | 242.5 TWh/half-year | 77.5 TWh/half-year | 165 TWh/half-year |
| 2026 | 277.4 TWh/half-year | 96.5 TWh/half-year | 180.9 TWh/half-year |
| 2027 | 317.3 TWh/half-year | 120.3 TWh/half-year | 197.1 TWh/half-year |
| 2028 | 363 TWh/half-year | 149.8 TWh/half-year | 213.2 TWh/half-year |
| 2029 | 415.2 TWh/half-year | 186.6 TWh/half-year | 228.6 TWh/half-year |
| 2030 | 475 TWh/half-year | 232.5 TWh/half-year | 242.5 TWh/half-year |
Water as half-year average
Billion liters divided by 2.
| Year | Total, bn L/half-year |
|---|---|
| 2023 | 280 bn L/half-year |
| 2024 | 312.2 bn L/half-year |
| 2025 | 348.1 bn L/half-year |
| 2026 | 388.2 bn L/half-year |
| 2027 | 432.8 bn L/half-year |
| 2028 | 482.6 bn L/half-year |
| 2029 | 538.1 bn L/half-year |
| 2030 | 600 bn L/half-year |
Absolute water mix per year
2023 shares on interpolated consumption.
| Year | Direct, bn L | Power, bn L | Hardware, bn L | Total, bn L |
|---|---|---|---|---|
| 2023 | 140 billion L | 373 billion L | 47 billion L | 560 billion L |
| 2024 | 156.1 billion L | 415.9 billion L | 52.4 billion L | 624.4 billion L |
| 2025 | 174.1 billion L | 463.7 billion L | 58.4 billion L | 696.2 billion L |
| 2026 | 194.1 billion L | 517.1 billion L | 65.2 billion L | 776.3 billion L |
| 2027 | 216.4 billion L | 576.6 billion L | 72.7 billion L | 865.6 billion L |
| 2028 | 241.3 billion L | 642.9 billion L | 81 billion L | 965.2 billion L |
| 2029 | 269.1 billion L | 716.8 billion L | 90.3 billion L | 1,076.2 billion L |
| 2030 | 300 billion L | 799.3 billion L | 100.7 billion L | 1,200 billion L |
MW stack per calendar year
Average MW from TWh and UTC seconds.
| Year | All, MW | AI, MW | Rest, MW |
|---|---|---|---|
| 2024 | 47,359 MW | 11,764 MW | 35,595 MW |
| 2025 | 55,365 MW | 17,694 MW | 37,671 MW |
| 2026 | 63,334 MW | 22,042 MW | 41,292 MW |
| 2027 | 72,449 MW | 27,458 MW | 44,990 MW |
| 2028 | 82,649 MW | 34,112 MW | 48,537 MW |
| 2029 | 94,803 MW | 42,611 MW | 52,192 MW |
| 2030 | 108,447 MW | 53,082 MW | 55,365 MW |
Leap years and UTC seconds
2024 has 366 days. TWh/s shifts slightly at the same CAGR.
| Year | Days | Seconds | Δ vs 365 days | All, TWh/s |
|---|---|---|---|---|
| 2024 | 366 | 31,622,400 | 86,400 | 0.00001316 TWh/s |
| 2025 | 365 | 31,536,000 | 0 | 0.00001538 TWh/s |
| 2026 | 365 | 31,536,000 | 0 | 0.00001759 TWh/s |
World vs IEA reference site, liters/day
2 million L/day illustration vs world daily mean (UTC days, interpolated between anchors).
| Year | World, L/day | Site, L/day | Sites at 2M L |
|---|---|---|---|
| 2024 | 1,706,049,733.3 L/day | 2,000,000 L/day | 853 |
| 2025 | 1,907,500,470.3 L/day | 2,000,000 L/day | 953.8 |
| 2026 | 2,126,911,401.2 L/day | 2,000,000 L/day | 1,063.5 |
| 2027 | 2,371,560,153.8 L/day | 2,000,000 L/day | 1,185.8 |
| 2028 | 2,637,124,718.3 L/day | 2,000,000 L/day | 1,318.6 |
| 2029 | 2,948,516,999.5 L/day | 2,000,000 L/day | 1,474.3 |
| 2030 | 3,287,671,232.9 L/day | 2,000,000 L/day | 1,643.8 |
Water reference values per AI TWh
Derived milliliters per request, scaled to one AI TWh. Google and OpenAI use different accounting boundaries. The values do not support a ranking.
| Year | ChatGPT, ml/AI TWh | Gemini, ml/AI TWh |
|---|---|---|
| 2024 | 946.4 billion | 1,083.3 billion |
| 2025 | 946.4 billion | 1,083.3 billion |
| 2026 | 946.4 billion | 1,083.3 billion |
| 2027 | 946.4 billion | 1,083.3 billion |
| 2028 | 946.4 billion | 1,083.3 billion |
| 2029 | 946.4 billion | 1,083.3 billion |
| 2030 | 946.4 billion | 1,083.3 billion |
OWID 945 vs IEA 950 (2030)
IEA Table A.1 versus the rounded executive summary in the same 2026 report. This is not a forecast revision.
| Metric | Value |
|---|---|
| OWID, TWh | 945 TWh |
| IEA, TWh | 950 TWh |
| Δ TWh | 5 TWh |
Reference calculation
Scale published figures to a chosen count
The calculator multiplies published figures by the count you enter. It does not measure your queries or estimate a personal footprint. ChatGPT uses OpenAI’s 2025 mean; Gemini uses the May 2025 median. IEA agent figures are modeled tasks using GPU electricity only. Training and your own device are excluded.
- ChatGPT, OpenAI 2025 (0.34 Wh)
- 6.8 Wh
- Gemini Apps, May 2025 (0.24 Wh)
- 4.8 Wh
- ChatGPT water (0.000085 gallons)
- 6.4 ml
- Gemini Apps, direct cooling water (0.26 ml)
- 5.2 ml
- IEA agent without reasoning (1.14 Wh)
- 22.8 Wh
- IEA agent with reasoning (50 Wh)
- 1,000 Wh
An example calculation of a published reference figure times the count entered. It is not a measured daily total or personal footprint.
IEA anchors
The annual figures behind the counter
Compact overview of the IEA year anchors for power and water. The wide detail table below is optional and shows derived scaling from the same numbers.
| Year | All data centers | AI-focused | Total water consumption |
|---|---|---|---|
| 2024 | 416 TWhIEA estimate | 103.3 TWhderived from +50% in 2025 | 624.4 billion Linterpolated |
| 2025 | 485 TWhIEA estimate | 155 TWhOWID reading of IEA figure | 696.2 billion Linterpolated |
| 2026 | 554.8 TWhinterpolated | 193.1 TWhinterpolated | 776.3 billion Linterpolated |
| 2030 | 950 TWhIEA base-case projection | 465 TWhIEA base-case projection | 1,200 billion LIEA base-case projection |
Detail table: MW, sites, households
Derived figures from the same anchors
| Year | Remainder without an AI focus | AI share | Average power, all | Average power, AI | 100 MW equivalents, all | 100 MW equivalents, AI | Household equivalents, all | Olympic pools, annual total |
|---|---|---|---|---|---|---|---|---|
| 2024 | 312.7 TWh | 24.8% | 47,359 MW | 11,764 MW | 473.6 | 117.6 | 47,358,834 | 199,813 |
| 2025 | 330 TWh | 32% | 55,365 MW | 17,694 MW | 553.7 | 176.9 | 55,365,297 | 222,796 |
| 2026 | 361.7 TWh | 34.8% | 63,334 MW | 22,042 MW | 633.3 | 220.4 | 63,333,575 | 248,423 |
| 2030 | 485 TWh | 48.9% | 108,447 MW | 53,082 MW | 1,084.5 | 530.8 | 108,447,489 | 384,000 |
Methodology
How the counter calculates
- 01
IEA annual anchors
Electricity comes from “Key Questions on Energy and AI” (April 16, 2026): around 416 TWh in 2024, 485 TWh in 2025, and around 950 TWh in 2030 for all data centers. Our World in Data reads IEA Figure 1.5 as 155 TWh for AI-focused sites in 2025. 155 plus 330 TWh of non-AI load equals the published 485 TWh. Those AI sites reach about 465 TWh by 2030. Table A.4 lists 416 TWh for 2024 and 945 TWh for the 2030 base case. For 2030, the counter follows the executive summary rounded to around 950 TWh. Water comes from “Energy and AI” (April 10, 2025): 560 billion liters consumed in 2023 and 1,200 billion liters in 2030.
- 02
Growth between anchors
Between two IEA years the annual total grows exponentially at the CAGR. That is why 2026 AI electricity sits near 193 TWh rather than a linear midpoint. After 2030 the counter holds the last IEA value. It does not invent its own long-range forecast.
- 03
Even spread across the calendar year
The IEA publishes annual totals, not hourly loads. Inside a UTC calendar year the counter spreads that total evenly across every second, including leap years. There is no seasonality, no outage, and no peak-shaving.
- 04
Water follows the electricity share
The IEA does not split water into AI and the rest. The counter applies the current AI share of data-center electricity to the global water total. The 2023 mix stays fixed. The IEA’s rounded shares imply approximately 25% direct cooling, 67% primary energy supply and electricity generation, and 8% hardware. Hardware stays out of the live ticker. The approximately 5,200 billion liters of 2023 withdrawals are independently read from IEA Figure 5.27, with graph-reading uncertainty. The ticker uses consumption only.
Formula
annualTotal(year) × (seconds since January 1 UTC / seconds in the year)
What this is not
Not a live meter, not a satellite wattmeter, not a real-time operator feed. The IEA figures are modeled too. Other models and accounting boundaries can produce different estimates. Cooling technology, climate, and the electricity mix affect water use. Keeping the water mix fixed through 2030 can overstate or understate later consumption.
Sources
Primary sources, not press copy
- IEA, Key Questions on Energy and AI
- IEA, Energy and AI
- Our World in Data, How much energy do data centers and artificial intelligence use?
- Google Cloud, How much energy does Google’s AI use? We did the math
- Sam Altman, The Gentle Singularity
- Epoch AI, How much energy does ChatGPT use?
- Bitkom/Borderstep, Rechenzentren in Deutschland
- CSO Ireland, Data Centres Metered Electricity Consumption 2025
- World Aquatics, Competition Regulations (February 2026)
- Microsoft, 2026 Environmental Sustainability Report (FY25)