Greenpeace report How Clean is Your Cloud? I saw mentioned in 3T magazine news is actually quite interesting reading. This year’s report provides a look at the energy choices some of the largest and fastest growing IT companies. The report analyzes the 14 IT companies and the electricity supply chain in more than 80 data center cases.
The report contains also lots of interesting background information on both IT and telecom energy consumption. I recommend checking it out. Here are some points picked from How Clean is Your Cloud? report:
Facebook, Amazon, Apple, Microsoft, Google, and Yahoo – these global brands and a host of other IT companies are rapidly and fundamentally transforming the way in which we work, communicate, watch movies or TV, listen to music, and share pictures through “the cloud.”
The growth and scale of investment in the cloud is truly mind-blowing, with estimates of a 50-fold increase in the amount of digital information by 2020 and nearly half a trillion in investment in the coming year, all to create and feed our desire for ubiquitous access to infinite information from our computers, phones and other mobile devices, instantly.
The engine that drives the cloud is the data center. Data centers are the factories of the 21st century information age, containing thousands of computers that store and manage our rapidly growing collection of data for consumption at a moment’s notice. Given the energy-intensive nature of maintaining the cloud, access to significant amounts of electricity is a key factor in decisions about where to build these data centers. Industry leaders estimate nearly $450bn US dollars is being spent annually on new data center space.
Since electricity plays a critical role in the cost structure of companies that use the cloud, there have been dramatic strides made in improving the energy efficiency design of the facilities and the thousands of computers that go inside. However, despite significant improvements in efficiency, the exponential growth in cloud computing far outstrips these energy savings.
How much energy is required to power the ever-expanding online world? What percentage of global greenhouse gas (GHG) emissions is attributable to the IT sector? Answers to these questions are very difficult to obtain with any degree of precision, partially due to the sector’s explosive growth, a wide range of devices and energy sources, and rapidly changing technology and business models. The estimates of the IT sector’s carbon footprint performed to date have varied widely in their methodology and scope. One of the most recognized estimates of the IT sector’s footprint was conducted as part of the 2008 SMART 2020 study, which established that the sector is responsible for 2% of global GHG emissions.
The combined electricity demand of the internet/cloud (data centers and telecommunications network) globally in 2007 was approximately 623bn kWh (if the cloud were a country, it would have the fifth largest electricity demand in the world). Based on current projections, the demand for electricity will more than triple to 1,973bn kWh (an amount greater than combined total demand of France, Germany, Canada and Brazil).
The report indicates that, due to the economic downturn and continued energy efficiency and performance improvements, global energy demand from data centers from 2005-2010 increased by 56%. Estimates of data center electricity demand come in at 31GW globally, with an increase of 19% in 2012 alone. At the same time global electricity consumption is otherwise essentially flat due to the global recession is still a staggering rate of growth.
Given the scale of predicted growth, the source of electricity must be factored into a meaningful definition of “green IT”. Energy efficiency alone will, at best, slow the growth of the sector’s footprint. The replacement of dirty sources of electricity with clean renewable sources is still the crucial missing link in the sector’s sustainability efforts according to the report.
The global telecoms sector is also growing rapidly. Rapid growth in use of smart phones and broadband mobile connections mean mobile data traffic in 2011 was eight times the size of the entire internet in 2000. It is estimated that global mobile data traffic grew 133% in 2011, with 597 petabytes of data sent by mobiles every month. In 2011, it is estimated that 6 billion people or 86.7% of the entire global population have mobile telephone subscriptions. By the end of 2012, the number of mobile connected devices is expected to exceed the global population. Electronic devices and the rapidly growing cloud that supports our demand for greater online access are clearly a significant force in driving global energy demand.
What about telecoms in the developing and newly industrialized countries? The report has some details from India (by the way it is expected that India will pass China to become the world’s largest mobile market in terms of subscriptions in 2012). Much of the growth in the Indian telecom sector is from India’s rural and semi-urban areas. By 2012, India is likely to have 200 million rural telecom connections at a penetration rate of 25%. Out of the existing 400,000 mobile towers, over 70% exist in rural and semi-urban areas where either grid-connected electricity is not available or the electricity supply is irregular. As a result, mobile towers and, increasingly, grid-connected towers in these areas rely on diesel generators to power their network operations. The consumption of diesel by the telecoms sector currently stands at a staggering 3bn liters annually, second only to the railways in India.
What is the case on other developing and newly industrialized countries? I don’t actually know.
NOTE: Please note that that many figures given on the report are just estimates based on quite little actual data, so they might be somewhat off the actual figures. Given the source of the report I would quess that if the figures are off, they are most probably off to direction so that the environmental effect looks bigger than it actually is.
943 Comments
Tomi Engdahl says:
There is an irony in using data center rage bait to get clicks on a story about agriculturally induced water stress considering the algo uses AI to feed the click bait to your readers.
Tomi Engdahl says:
Here are the 4 cooling systems available for data centers. Essentially lowest to highest cost as listed.
1. Evaporative Cooling: Water absorbs heat and then evaporates in cooling towers. While highly energy-efficient, this method continuously consumes water.
2. Closed-Loop Systems: Water (or a water-glycol mix) circulates in a sealed system to carry heat away, similar to a car radiator. This drastically reduces or eliminates daily water consumption, though it may require more electricity to run mechanical chillers.
3. Outside Air (Free Cooling): In colder, temperate regions, data centers pull in filtered outdoor air to naturally cool the facility, relying on water only on the hottest days.
4. Direct-to-Chip & Immersion Cooling: Specialized liquids or water are pumped directly to microprocessors or servers submerged in dielectric fluid, isolating the delicate hardware from direct water contact.
Tomi Engdahl says:
1 and 2 arent mutually exclusive. 4 is one of the other 3. You have wet chillers, dry chillers, or standard RTUs. Wet or dry chillers are closed loop systems and both can either be direct to chip or air cooled.
Tomi Engdahl says:
Breakdown by Sector
A. Agriculture (~70% Global / ~47% U.S. Withdrawals)
B. Industrial & Energy (~20% Global / ~40% U.S. Withdrawals)
C. Domestic & Municipal (~10% Global / ~13% U.S. Withdrawals)
Data centers:
~0.15% of total U.S. water consumption
~0.004% to 0.05% of worldwide freshwater
Tomi Engdahl says:
“Once water is used, it is contaminated.”
The localized burden on freshwater reserves remains one of the largest criticism points of the digital infrastructure boom.
In standard evaporative cooling towers, water never touches electronic circuit boards or heavy metals. It circulates through heat exchangers to absorb heat from the air or closed coolant loops, then evaporates as steam into the atmosphere.
Around 80% to 90% of the water drawn for evaporative cooling is lost to the air as pure steam/vapor (which contains zero contaminants, leaving the local watershed entirely).
The remaining 10% to 20% of the water becomes heavily concentrated with naturally occurring minerals (like calcium and magnesium) left behind by evaporation, along with anti-scaling, anti-corrosion, or anti-microbial additives (like chlorine or biocides).
Data centers are not legally permitted to inject dirty cooling wastewater back into underground aquifers.
Blowdown water is sent directly to municipal wastewater treatment facilities or treated on-site before being discharged into surface waterways under strict Environmental Protection Agency (EPA) National Pollutant Discharge Elimination System (NPDES) permits.
Direct Groundwater Risks: The main risk of contamination to groundwater wells comes from secondary facility sources—such as potential leaks or spills from massive on-site diesel generator fuel tanks used for backup power—rather than the cooling water itself.
Tomi Engdahl says:
https://www.facebook.com/share/p/1BTdFG4ZfZ/
A large share of new artificial intelligence (AI) data centers planned across the United States are expected to be built in regions already experiencing drought and water shortages, according to an analysis, the expansion of AI infrastructure is increasing concerns about the pressure these facilities could place on limited natural resources, especially in dry areas.
AI data centers require enormous amounts of electricity and water to operate. These facilities contain thousands of computer servers that process and store data for artificial intelligence systems, cloud services, and digital applications. Because servers generate significant heat, many data centers rely on cooling systems that often require large quantities of water.
The analysis found that many proposed AI data centers are being located in regions where water availability is already under stress. Some projects are planned in areas affected by long-term drought conditions, raising concerns among local communities, environmental groups, and researchers about future competition for water resources.
One example is the proposed Stratos Project in Box Elder County, Utah, where developers plan to construct a massive data center campus. The project covers around 40,000 acres and could require up to 9 gigawatts of electricity. The area has faced challenges related to water availability, making the scale of the project a topic of debate.
The rapid growth of artificial intelligence has led technology companies to invest heavily in new data centers. These facilities are essential for training and running advanced AI models, but their environmental impact has become a growing concern. Beyond water consumption, they also require large amounts of energy, which can increase demand on regional power grids.
Companies developing data centers have argued that newer technologies can reduce resource usage. Some facilities are adopting more efficient cooling systems, renewable energy sources, and improved designs to lower their environmental footprint. However, critics say that the overall increase in the number and size of AI data centers could continue to create significant demands on local resources.
Communities near planned projects have raised questions about whether the benefits of new technology investments outweigh the potential environmental costs. While data centers can bring economic opportunities, including jobs and infrastructure development, residents in drought-prone regions are concerned about how increased water and energy use could affect agriculture, ecosystems, and future water supplies.
The situation highlights a growing challenge as artificial intelligence continues to expand. The technology requires physical infrastructure that depends on real-world resources, including land, electricity, and water. As companies race to build more AI facilities, policymakers and communities are increasingly focused on finding a balance between technological progress and environmental sustainability.
The debate over AI data centers reflects a broader question about the future of digital growth: how can society continue advancing artificial intelligence while ensuring that essential natural resources remain protected for the communities and environments that depend on them?
Tomi Engdahl says:
https://www.facebook.com/share/p/1BSTqqoa5k/
Global data centers will consume unprecedented amounts of power, according to a landmark report from the United Nations University. This rapid growth is driven primarily by artificial intelligence infrastructure, which represents the fastest-growing sector within global data center operations.
According to researchers from the UN University Institute for Water, Environment and Health, global data centers are projected to consume 945 terawatt-hours of electricity annually by 2030. This represents a near-doubling from 2025 consumption levels and exceeds the combined annual electricity use of Pakistan, Bangladesh, and Nigeria—countries home to more than 650 million people.
However, researchers warn that focusing solely on carbon emissions is a critical mistake. Every data center depends on massive physical infrastructure demanding vast tracts of land, raw materials, and enormous quantities of water for cooling. By 2030, the water footprint associated with global data center operations could equal the basic annual household water needs of all 1.3 billion people in Sub-Saharan Africa, while the land footprint could exceed 14,500 square kilometers, roughly twice the area of metropolitan Jakarta.
Crucially, the report highlights that 80 to 90 percent of AI’s electricity use occurs during inference—the process of answering everyday user queries—rather than the initial training phase. This means operational usage now dominates energy consumption. The energy demands vary dramatically by task: a typical chat query uses roughly 200 times the energy of basic text classification, while generating a single AI image consumes approximately 1,450 times more energy than text classification.
Researchers urge policymakers to adopt a holistic framework that measures the complete material footprint, including water, land, and electronic waste, not just carbon emissions.
Images are generated by AI and for demonstration purposes only.
Source: Aczel, M., Chamanara, S., Matin, M., Farsi, A., Marwala, T., & Madani, K. (2026). *Environmental Cost of AI’s Energy Use: Carbon, Water and Land Footprints.* United Nations University Institute for Water, Environment and Health (UNU-INWEH).
#sustainability #energy #technology #ai #environment #fblifestyle
Tomi Engdahl says:
“All you need to know is how much heat those servers produce, then you’ll know how much energy is needed, at a minimum, to cool them.”
That information is just a starting point for estimation how much energy is needed for cooling. Cooling method and environment affects considerably needed energy for cooling. On average across the global data center fleet, computing takes roughly 60% to 75% of total facility energy, while cooling consumes about 15% to 40%.
Legacy / Traditional Enterprise Facilities (PUE approx 1.5 to 1.8).
The Ratio: For every 1.0 kW used by the servers, an additional 0.5 to 0.8 kW is wasted on cooling chillers, fans, and pumps.
Modern Hyperscale Facilities (e.g., Google, Microsoft, Meta) (PUE approx 1.1 to 1.2).
The Ratio: For every 1.0 kW used by the servers, only about 0.1 to 0.2 kW goes toward cooling. These facilities achieve this through free-air economizers, hot/cold aisle containment, and optimized ambient operating temperatures.
“If there’s a system that can cool that much heat without needing fluid refills or meaningful amounts of electricity, I want one for my house!”
There is. But it does not work everywhere. Build data center to cold climate, use cold outdoor air and reuse waste heat for district heating.
Tomi Engdahl says:
In high-density air-cooled racks, internal server fans can draw up to 10% to 15% of the server’s own electrical draw.
Switching to liquid immersion cooling removes the need for chassis fans entirely, which actually lowers the IT compute load itself while shrinking overall facility cooling draw.
Submerging computers make transferring the heat from computer boards to data center cooling system more efficient, but the heat used by computer chios need still go to somewhere (chillers) or otherwise after few hours we have a lot of hot non conductive fluid.
Tomi Engdahl says:
The biggest advantage of a closed-loop liquid system (especially direct-to-chip or immersion cooling) is that it captures server heat at much higher temperatures than air systems—often pushing coolant out at 50°C to 70°C+.
Because the loop is closed, that heat isn’t dumped into the atmosphere. Instead, it runs through a brazed plate heat exchanger (BPHE) to transfer its energy into a secondary loop:
District Heating: The heat is piped directly into municipal heating grids to warm nearby homes, offices, or greenhouses.
Heat Pumps: Industrial heat pumps can raise that closed-loop water from 50°C to 80°C+ with minimal energy, making it ideal for municipal hot water networks.
Tomi Engdahl says:
evaporative seawater cooling systems exist, but they are engineered differently than standard freshwater cooling towers.
While seawater is abundant, evaporating salt water creates severe engineering challenges.
Direct Seawater Evaporative Towers
Closed-Loop / Wet Surface Air Coolers (WSAC)
The Non-Evaporative Alternative: SWAC (Seawater Air Conditioning)
While direct evaporative seawater cooling is technically possible, the severe corrosion, scaling, and salt-drift management usually make it less desirable than closed-loop indirect evaporative systems (WSAC) or non-evaporative deep seawater heat exchange (SWAC).
Tomi Engdahl says:
Facebook claims:
Perpetual moratorium on all data centers.
It’s been scientifically proven data centers are detrimental to humans, animals, and fish.
In addition, the massive amounts of water needed for cooling the centers depletes ground water, lakes, streams, rivers, an oceans.
After the water has circulated through the data center it is contaminated with heavy metals that cause still births, deformities, retardation, and death to unborn and small children.
Cancer is rampent in areas that have data centers.
Answers:
This claim blends real environmental concerns with severe, hyperbolic misrepresentations. While data centers do present genuine ecological and public health challenges—particularly regarding regional water stress and indirect air pollution from power grids—claims of guaranteed “rampant cancer,” toxic heavy metal discharges causing severe birth defects, or ocean depletion are scientifically ungrounded exaggerations.
The statement uses sensationalist language (“perpetual moratorium,” “rampant cancer,” “depleting oceans”) to turn real industrial challenges into an apocalyptic scenario.
data centers do demand rigorous regulation regarding clean energy adoption, closed-loop water systems, and local air quality controls, so that their actual environmental footprint would be manageable engineering and policy challenge.
Claim 1: “It’s been scientifically proven data centers are detrimental to humans, animals, and fish.”
The Reality: Partially true regarding regional impacts, but false as a blanket absolute.
Claim 2: “Water needed for cooling depletes groundwater, lakes, streams, rivers, and oceans.”
The Reality: Significant for local aquifers; absurd regarding oceans.
Groundwater & Lakes: This is a legitimate issue in water-stressed regions. In places like Arizona, Texas, or parts of Oregon, a single large data center can consume millions of gallons of fresh water daily for evaporative cooling, placing severe strain on local aquifers and drinking water supplies shared with local communities.
Oceans: It is physically impossible for data centers to “deplete oceans.”
Claim 3: “Water is contaminated with heavy metals causing stillbirths, deformities, retardation, and death…”
The Reality: Highly misleading; conflating e-waste recycling with server operation.
Cooling Water Discharge: Data center wastewater (called “blowdown”) can carry concentrated minerals, algae-control biocides, and trace heavy metals (like copper) leached from internal plumbing. However, this water is heavily regulated under industrial wastewater laws and must be treated before discharge.
Where the Extreme Claims Come From: The extreme health effects described (stillbirths, severe birth defects, and neurological damage) are scientifically linked to informal electronic waste (e-waste) recycling—such as open-air burning of discarded circuit boards and melting lead/cadmium—not the operational cooling water coming out of a data center building.
The Nitrate Concentration Issue: In specific cases (like a publicized dispute in Oregon), data centers using groundwater already polluted by agricultural runoff evaporated pure water, leaving behind higher concentrations of agricultural nitrates in the wastewater. The data center didn’t create the toxic heavy metals or nitrates, but its evaporation process concentrated existing agricultural pollutants.
Claim 4: “Cancer is rampant in areas that have data centers.”
The Reality: False statement of direct causation.
Air Quality & Risk: There is no scientific evidence of “rampant cancer clusters” caused directly by living near a data center building. However, long-term exposure to diesel particulate matter from backup generators or nearby fossil-fuel power plants does incrementally elevate baseline cancer and respiratory risks over decades—a risk shared by any community near heavy industrial zones, highways, or power plants.
Zoning Correlation: Studies tracking public health near data center clusters (like California’s CalEnviroScreen data) note that data centers are frequently built in industrial zones that already had higher pre-existing pollution levels from trucking hubs, factories, and agricultural runoff.
Tomi Engdahl says:
Source: https://www.facebook.com/share/p/1EgDnqP3iS/
Tomi Engdahl says:
OUR TECHNOLOGY MUST BALANCE WITH OUR PLANET.
The massive expansion of AI data centers is reshaping our world. We need to set firm guardrails right now so future generations don’t pay the price for unchecked industrial growth.
We are proposing the CAFABI Mandate Act—a new framework that legally forces heavy digital infrastructure to harmoniously adapt to local ecosystems.
What does CAFABI stand for?
Continuous — Real-time tracking, eliminating useless annual checkups.
Adaptive — Technology that shifts its power output based on the environment’s needs
Feedback — Automated alerts that trigger instant corrections.
Assessment — Scientific verification of ecosystem health.
Balanced Infrastructure — Building tech hubs that actively protect the earth.
How it works:
Live IoT Monitoring: Constant, real-time data feeds track local ecosystem strain
Species Protection: Enforces strict limits on noise, heat, and light to protect the natural life cycles of local wildlife, insects, and amphibians.
Saving Critical Resources: Mandates strict energy efficiency and bans data centers from using public drinking water for cooling.
Automated Enforcement: If a facility breaks environmental safety thresholds, secure software automatically issues fines or throttles their power until balance is restored.Let’s build the future right the first time.
What do you think? Should data centers be legally required to plug into an automated environmental protection loop? Let us know below!
#CAFABIMandate #DataCenters #GreenTech #Sustainability #EcoFriendly #Infrastructure
https://www.facebook.com/share/p/1EgDnqP3iS/
Tomi Engdahl says:
https://www.facebook.com/share/p/18v2vD3cgt/
The pollution generated by AI data centers has escalated into a crisis of staggering proportions, consuming vast quantities of electricity and water while emitting greenhouse gases at scales that strain global resources.
These facilities, powering the explosive growth of artificial intelligence, demand enormous energy inputs—often rivaling the consumption of entire nations.
Global data centers already use hundreds of terawatt-hours annually, producing carbon dioxide emissions comparable to major countries like Argentina, with projections indicating a near doubling in the coming years as AI adoption surges.
This energy hunger frequently relies on fossil fuel-heavy grids, releasing millions of metric tons of CO2 equivalent, while cooling systems devour billions of liters of water, exacerbating shortages in stressed regions and competing with agriculture and households.
Beyond direct emissions, the environmental toll includes land use for infrastructure, electronic waste from rapid hardware obsolescence, and indirect pollution from mining rare materials for servers and chips.
The cumulative impact risks undermining net-zero goals, as unchecked expansion could equate to adding millions of cars to roads or consuming water for millions of residents.
What makes it nearly incomprehensible is the speed: an industry once marginal now rivals aviation or heavy industry in footprint, hidden behind sleek digital interfaces yet profoundly altering planetary boundaries.
Source: United Nations University Report: “Environmental Cost of Artificial Intelligence: Carbon, Water, and Land Footprints” (2026).
Tomi Engdahl says:
Water Quality: Potable vs. Raw Water
This is often where the policy conflict arises:
Corn Irrigation: Primarily relies on raw groundwater aquifers (like the High Plains/Ogallala Aquifer), surface river diversions, or untreated non-potable agricultural water.
Data Centers: Most evaporative data center cooling systems require high-quality, treated potable (drinking) water from municipal utility networks to avoid scaling, mineral buildup, and biological growth on cooling equipment.
Tomi Engdahl says:
Using desalinated water for data center cooling introduces a direct trade-off between mitigating local freshwater depletion and incurring high energy, financial, and environmental costs.
Rather than desalinating seawater to feed an evaporative tower, data center engineers increasingly compare desalination against alternative architectures:
A. Closed-Loop Air/Dry Coolers (Water-Free)
B. Direct Seawater Heat Exchange (SWAC / Wet Surface Cooling)
C. Waste-Heat Driven Desalination (Thermally Integrated)
In the cooling hierarchy, desalination is generally considered a last-resort water source. Operators prefer reclaimed wastewater first for evaporative systems because it requires less energy to treat than ocean brine.
For coastal facilities looking to bypass freshwater entirely, closed-loop direct ocean heat exchangers or closed-loop dry coolers are far more economically and environmentally viable than building on-site desalination infrastructure.
Tomi Engdahl says:
https://www.facebook.com/share/p/1J8iG6PqZq/
Britain reportedly wants to triple its data centre capacity by 2030, but according to Water UK, the trade body representing the country’s water companies, current planning may not account for enough water to support that growth
In written evidence submitted to Parliament, Water UK said government forecasts appear to leave data centres out of long-term water planning entirely.
Data centres in England are said to use around 6.6 million litres of water daily, a figure that could reportedly approach 20 million litres if capacity triples as proposed. Many of these facilities are located in regions already dealing with water stress and hosepipe restrictions. A House of Lords report has suggested England could face a significant daily shortfall in public water supply by the year 2055 if current trends continue. Experts say this raises real questions about how growth and resource planning intersect.
Sources: Water UK Parliamentary submission, House of Lords report, AOL News.
Tomi Engdahl says:
Going green:
Best case scenario – we save the planet
Worst case scenario – we unintentionally reduce the need for foreign energy
Tomi Engdahl says:
Data Centers
Evaporation (70% to 80%): In standard cooling towers, heat from the servers warms the water, causing a large majority of it to evaporate into the air as water vapor. This vapor enters the atmosphere and rejoins the global water cycle, though it is permanently lost from the local immediate watershed. [1, 2, 3, 4]
Wastewater / Blowdown (20% to 30%): The remaining water that does not evaporate is drained from the system as warm liquid effluent. This water often contains chemical additives (like anti-corrosives or biocides) used to maintain the cooling equipment. It is typically sent to a local municipal wastewater treatment plant before being released or reused downstream. [1, 2, 3]
Nuclear reactors operate at core temperatures ranging from 300 °C (570 °F) in standard water-cooled plants to over 950 °C (1,740 °F) in high-temperature gas designs. By contrast, experimental nuclear fusion reactors can reach extreme temperatures of 150 million °C, which is hotter than the center of the sun
Coal power plant furnaces burn at around 2,400°F to 2,700°F (1,315°C to 1,480°C), creating high-pressure steam inside the boiler pipes that reaches roughly 1,000°F to 1,200°F (538°C to 649°C). [1, 2, 3, 4, 5]
Furnace and Flame Temperatures
* Combustion chamber: The burning pulverized coal reaches peaks between 2,400°F and 2,700°F
*
Data center operating floors are typically kept between 64.4°F and 81°F (18°C to 27°C) to protect hardware. However, the individual servers and AI chips inside the facility run much hotter, often using liquid cooling that reaches 113°F (45°C) or higher to safely pull heat away from the hardware. [1, 2, 3, 4, 5]
Because they consume massive amounts of power, modern data centers expel huge quantities of waste heat into the environment.
We have been cooling power plants for decades which run much hotter than data centers and we don’t see evaporation rates as high as what we are being told is going to happen with data centers. What are there really doing with the water. Soon water is going to be scarce because of the data centers which produces heat much lower than power plants. This makes me think they the ones in power that have world wide agendas are now creating a scarcity of water and now we are going to pay. The equations between coal nuclear heat coal heat and other fuel heats is not matching the evaporation. Where is the water really going ?
Source: comment https://www.facebook.com/share/v/18J6zNJPfC/
Tomi Engdahl says:
Instead of building entirely new industrial sites, Finland is repurposing former paper mill locations for clean-energy projects. These sites already have valuable infrastructure, including high-capacity electricity connections, water supplies, rail links, roads, and industrial zoning, making them ideal for green hydrogen production. Reusing existing facilities can reduce construction costs, shorten project timelines, and breathe new life into regions affected by paper mill closures. The approach supports Finland’s ambition to become a leading producer of green hydrogen while cutting carbon emissions and creating new jobs through sustainable industrial redevelopment.
#Factify #GreenHydrogen #finland #CleanEnergy
Tomi Engdahl says:
Comparing Direct-to-Chip (DTC) liquid cooling—the dominant land-based solution for high-density AI clusters—with submerged/underwater data center cooling highlights two fundamentally different engineering approaches.
While DTC optimizes heat removal at the rack level inside standard warehouses, underwater systems turn the entire ocean or sea bed into a massive, natural heat sink.
Tomi Engdahl says:
https://www.facebook.com/share/p/1BbRuBQvtp/
The hottest part of AI right now is not the algorithms. It is the heat.
As artificial intelligence infrastructure expands around the world, cooling has become one of the industry’s biggest sustainability headaches. Traditional air-cooled facilities can demand enormous amounts of water and electricity, which raises environmental concerns as AI use accelerates.
Modern AI training and inference pack high-performance chips into dense racks. That density generates intense heat, and moving it away with air alone gets harder and less efficient at scale.
Enter advanced liquid cooling. By circulating specialized fluids directly over processors, these systems pull heat away far more efficiently than conventional air methods. The result is a dramatic reduction in water consumption and overall energy use, especially important as more data centers are built in regions already facing tight resource constraints.
There is another upside. Liquid cooling can make waste heat recovery practical, so the warmth that once disappeared into the atmosphere can be captured and put to work. Depending on local infrastructure, that energy could help feed nearby heating needs or support district energy systems, shifting data centers from pure resource consumers toward contributors.
Communities and planners are watching this closely because cooling choices affect local water use, grid demand, and where facilities can responsibly be sited. As computing needs grow, smarter thermal management is likely to be a cornerstone of sustainable digital infrastructure.
If your city invited a liquid-cooled data center that agreed to reuse its waste heat locally, would you support it?
#SustainableComputing #DataCenterEfficiency #AIInfrastructure
Tomi Engdahl says:
https://www.fortum.com/fi/lammitys-ja-jaahdytys/tuotantomme/espoo-clean-heat/datakeskusten-hukkalammon-kierratys
Tomi Engdahl says:
When comparing the heat generated by global data centers to the excess heat absorbed by the oceans, the numbers reveal an astounding difference in physical scale.
While data centers are major consumers of electricity on land, their total heat generation is a drop in the ocean compared to the climate-driven thermal energy added to the seas each year.
The Scale Ratio: The world’s oceans absorb approximately 15,000 times more heat every year from greenhouse-gas-driven global warming than the direct heat output generated by every data center on Earth combined.
Energy Comparison: The ~23 Zettajoules absorbed by the oceans in a single year is equivalent to roughly 200 times humanity’s total annual electricity consumption across all homes, factories, vehicles, and data centers.
The Real Impact (Grid Carbon): The actual climate impact of data centers isn’t the heat radiating from their server chips, but rather the carbon emissions produced if the electricity powering those servers comes from fossil fuels. Those greenhouse emissions trap solar energy in the atmosphere, which then warms the oceans over time.
Tomi Engdahl says:
Summary
If every data center on Earth were plunged directly into the ocean to dump its waste heat into the water, the direct temperature rise of the global ocean would be unmeasurable.
The thermal challenge of data centers is purely local (managing localized water usage, regional power grid capacity, and local hot-air exhaust), whereas ocean warming is a global atmospheric phenomenon driven by planetary greenhouse gas accumulation.
Tomi Engdahl says:
Placing a cluster of high-voltage, high-capacity computing hardware directly into marine environments introduces specific ecological stressors beyond thermal dissipation:
1. Acoustic Pollution (Low-Frequency Noise)
2. Electromagnetic Fields (EMF)
3. Seafloor Benthic Disturbance & Habitat Destruction
4. Biofouling Control & Chemical Leaching
5. Localized Micro-Habitat Shifts (The “Attraction” Paradox)
Heat is the most measurable thermodynamic output, but acoustic hum, electromagnetic interference, biofouling mitigation, and physical seabed disruption represent the primary ecological hurdles that complicate widespread subsea deployments.
Tomi Engdahl says:
there are data centers that use lots of water and data centers that use very little water. In USA it has been possible and more economical for companies to build and run data centers that use lots of water.
Tomi Engdahl says:
Roughly 40% of all U.S. data centers sit in areas experiencing high or extreme water stress (such as Phoenix, Arizona, and North Texas), where evaporative cooling directly competes with residential drinking water during summer droughts.
Tomi Engdahl says:
However, because nearly two-thirds of them plug directly into local municipal drinking water lines in hot, dry, or dense suburban areas, they can suddenly account for 5% to 25%+ of a single city’s public water supply.
Tomi Engdahl says:
https://www.facebook.com/share/1H6dzZoF1q/
Erin Brockovich, the environmental activist who helped expose one of America’s biggest corporate pollution scandals, has now turned her attention to the rapid expansion of AI data centres.
In the 1990s, Brockovich uncovered evidence that Pacific Gas and Electric (PG&E) had contaminated groundwater in Hinkley, California, with toxic hexavalent chromium. Her investigation helped lead to a landmark $333 million settlement in 1996 for more than 600 residents whose lives had been affected.
Now, three decades later, she is investigating the massive facilities being built to power the AI revolution.
In April 2026, Brockovich launched a public map and reporting platform, encouraging people to document AI data centres that are being built or proposed near their communities.
Since then, more than 8,200 reports have been submitted, with residents raising concerns about heavy water use, increasing electricity costs, constant noise, pollution, impacts on wildlife, and projects moving forward without communities fully understanding their environmental effects.
Brockovich has not claimed that every data centre is harmful, nor has she announced a sweeping lawsuit against the AI industry.
Instead, she says she is collecting evidence, identifying patterns, and asking difficult questions about who benefits from the AI boom—and who bears its environmental costs.
Thirty years ago, her persistence helped uncover one of the most famous environmental scandals in American history.
Today, she’s asking whether the race to build the infrastructure behind artificial intelligence could be creating another.
Tomi Engdahl says:
Demand from AI data centers is straining electric grids and forcing companies to find new ways to power their facilities.
Read more: https://kdvr.com/news/nationalworld-news/data-centers-to-use-1-5-of-us-power-by-2035-report/
Tomi Engdahl says:
While putting solar panels on the flat roofs of massive data centers seems like a no-brainer, three physical and structural realities explain why rooftop solar provides surprisingly little relief:
1. The Energy Density Mismatch
The Math: A commercial solar panel generates roughly 15 to 20 watts per square foot during peak daylight. Even if you cover every single square inch of a data center roof with solar panels, it generates less than 2% to 5% of the power the servers inside consume.
2. The Roof Is Already Full
Adding solar panels means squeezing them into the gaps between hot air exhausts, which lowers panel efficiency due to thermal heat, or structurally reinforcing the roof to elevate panels above heavy equipment.
3. Intermittency vs. 24/7/365 Demand
Data centers cannot throttle down when a cloud passes overhead or when the sun sets at night; they require continuous, high-density baseload power every second of the day. Rooftop solar generates peak energy for about 4 to 6 hours a day and zero at night, requiring massive battery banks if used directly off-grid.
Tomi Engdahl says:
The shift toward zero-water or ultra-low water data centers in the U.S. is already underway—mostly driven by local drought restrictions, public pushback, and environmental regulations in dry states like Arizona, Texas, and Utah.
A data center can achieve near-zero or zero water consumption using three primary engineering strategies:
1. Closed-Loop Dry Cooling & Air-Cooled Facilities
Instead of evaporating water in cooling towers to reject heat into the atmosphere, these facilities function like a massive car radiator. A sealed fluid (often water blended with glycol or synthetic refrigerant) circulates in a closed loop, dissipating server heat into the ambient air via external fans. Water Usage: Near Zero (water is only added once during installation to fill the closed loop, with zero daily evaporation).
2. Direct-to-Chip & Immersion Cooling (Dielectric Fluids)
Synthetic fluids (like fluorinated liquids or synthetic oils) absorb heat directly from high-power chips and transfer it to dry outdoor heat exchangers without relying on municipal water lines. Water Usage: Zero Water.
3. Industrial Reclaimed / Recycled Water Facilities
While not technically “waterless,” some major data center hubs completely eliminate their draw on local drinking water (potable municipal supplies) by building dedicated purple-pipe infrastructure to run on treated industrial wastewater or greywater. Water Usage: Zero fresh or drinking water.
Tomi Engdahl says:
For water usage there are both thirst and almost no water needed data center in use and being built to USA.
Microsoft mandates zero-water-evaporation designs for all new datacenter builds.
Meta shifted its baseline design across new facilities (such as in Kuna, Idaho, and Mesa, Arizona) to use direct outdoor air cooling combined with closed-loop liquid systems, avoiding evaporative water towers for the vast majority of the year.
Tomi Engdahl says:
at the moment data centers use around 0.2% of drinkable water in USA. At some locations where water was already in short supply the decision makers for some reason allowed to build data centers that use lots of water.
Typical situation:
1. In many areas there has been water short supply and over-use even before data centers
2. Adding data centers made the water shortage problem worse (more or less)
If you would close down data centers, it would somewhat help the water shortage, but would not fix the situation back to good.
The problem is in many places deeper than blaming data centers: history of miscalculation of available water resources and changing climate (less rain to certain areas).
To bring the water usage at current climate to sustainable level would need scaling back agriculture, industry and data centers so that they use considerably less water.
Tomi Engdahl says:
Data centers are not a new idea. The idea that data centers originated in the 1940s. After that the computing trends have over time what is optimal between between centralized data centers and distributed computing.
Tomi Engdahl says:
Do we still need to conserve on water or electricity to save the planet ??!
Depends on your location and how your electricity is made. If your area is using more water than sustainable, conserving water there is a good idea. If water use is already sustainable where you are, reducing existing water use does not help you or those that lack water.
The trend now in Europe is to reduce use of fossil fuels (co2 reduction) and increase use of “clean” / “co2 free” electricity (sun,wind,water and in some cases nuclear)
Tomi Engdahl says:
https://www.facebook.com/share/p/14mPa8TgiCp/
The United States now accounts for nearly 40% of the world’s data center electricity consumption, making it the largest consumer by a wide margin.
Recent energy data shows that U.S. data centers used 312.6 terawatt-hours (TWh) of electricity in a single year—about 52% more than China, the world’s second-largest user. That amount of power is comparable to the annual electricity demand of roughly 29 million American homes, highlighting how rapidly AI and cloud computing are reshaping energy consumption.
The surge has been driven largely by the AI boom, with technology companies investing heavily in new computing infrastructure. Between 2022 and 2025, global AI-related spending climbed from $162 billion to $448 billion, fueling a sharp increase in data center construction and electricity demand.
Energy analysts warn that data centers now consume around 6.6% of all electricity generated in the United States. While overall U.S. power generation has increased only modestly in recent years, electricity use by data centers has surged by more than 80%, raising concerns that grid expansion and new energy sources will need to keep pace with the country’s rapidly growing AI industry.
Source: Climate Crisis 247 (2026), citing the Energy Institute’s Statistical Review of World Energy (2026).
Tomi Engdahl says:
https://www.fwpcoa.org/content.aspx?page_id=5&club_id=859275&item_id=130961
News / Articles
Myths vs. Reality: Data Centers And Water Usage
Tomi Engdahl says:
Water consumption vs. water withdrawal: It’s crucial to distinguish these terms. When a data center “withdraws” water, that’s the amount taken from the source (e.g. pumping from the city water line). “Consumption” means water that’s actually used up (not returned) – primarily through evaporation. The difference (withdrawal minus consumption) is the water returned, usually as wastewater (warm water or “blowdown” drained from cooling systems). In a data center cooling context, most of the water that is withdrawn ends up consumed. Typically, 70–80% of the water in evaporative cooling is lost as evaporation into the air. The remaining 20–30% is discharged as liquid wastewater (which goes to a sewer or treatment plant).
https://www.fwpcoa.org/content.aspx?page_id=5&club_id=859275&item_id=130961
Tomi Engdahl says:
https://genesiswatertech.com/blog-post/treated-wastewater-for-data-center-cooling-a-practical-guide-to-alternative-water-sources/
Wastewater Reuse for Data Center Cooling: A Practical Guide
Data centers consume approximately 200 billion gallons of water annually in the United States alone, with cooling systems accounting for 80-90% of that demand. As water scarcity intensifies across major data center markets—from Phoenix to Northern Virginia—operations managers face a critical challenge: maintaining uptime while addressing water availability constraints and rising costs.
Wastewater reuse for data center cooling has emerged as the most viable solution to this challenge. Forward-thinking facilities are already proving that municipal, industrial, and on-site wastewater can meet cooling system requirements while delivering measurable operational and sustainability benefits through advanced data center water treatment.
Water Scarcity: The Business Case for Alternative Sources
Water stress is no longer a future concern—it’s impacting data center operations today. The World Resources Institute classifies 17 U.S. states as experiencing high or extremely high water stress. These states host approximately 40% of U.S. data center capacity.
Three market forces are accelerating the shift toward alternative water sources for data centers:
Regulatory restrictions: Municipalities in water-stressed regions are implementing allocation limits on potable water for industrial cooling.
Cost escalation: Potable water rates for commercial users have increased 43% over the past decade in major data center markets. Facilities consuming 50-100 million gallons or larger annually face water costs exceeding $500,000 yearly—before factoring in wastewater discharge fees.
Stakeholder expectations: Corporate water stewardship commitments and ESG reporting requirements have made water intensity a key performance indicator. Hyperscale operators and colocation providers are setting targets for water neutrality or positive water impact, driving demand for alternative water sources in data centers.
Treated wastewater typically costs 30-50% less than potable water in most markets.
Not all wastewater is equal for cooling applications. Three primary source categories offer distinct advantages and implementation considerations for water reuse data centers:
Municipal wastewater: Treated effluent from municipal wastewater treatment plants represents the most scalable option. These facilities produce consistent volumes with relatively predictable quality characteristics. Municipal sources work best for data centers located within 5-10 miles of treatment plants, minimizing conveyance infrastructure costs.
Industrial wastewater: Manufacturing facilities, refineries, and power plants generate significant wastewater volumes. On-site treatment or collaborative treatment arrangements can convert these streams into cooling water supplies. Industrial wastewater partnerships work particularly well for data centers co-located near industrial parks or manufacturing zones.
Quality varies significantly based on source operations. Food and beverage industry wastewater contains high organic loads requiring biological treatment. Semiconductor manufacturing wastewater may contain specific contaminants but typically features lower total dissolved solids than municipal sources.
On-site wastewater: Data centers generate wastewater from cooling tower blowdown, reverse osmosis reject streams, and facility processes. Advanced data center water treatment systems can recover 70-90% of this water for reuse in cooling systems.
Data Center Water Treatment Requirements for Cooling Applications
Cooling systems impose specific water quality requirements to prevent scaling, corrosion, biological growth, and fouling. Data center water treatment design must address both source water characteristics and end-use specifications.
Key water quality parameters for cooling applications include:
Total dissolved solids (TDS): Target range of 500-1,500 mg/L for most cooling tower systems. Higher TDS concentrations increase scaling potential and corrosion rates. Municipal wastewater typically ranges from 600-1,200 mg/L TDS after secondary treatment.
Hardness: Calcium and magnesium concentrations should remain below 200-400 mg/L as CaCO₃ to minimize scale formation. Softening or scale inhibitor programs manage hardness in source water exceeding these levels.
Suspended solids: Cooling water specifications typically limit suspended solids to 10-25 mg/L. Particulates cause fouling in heat exchangers and promote microbial growth in cooling towers.
Biological content: Total coliform and heterotrophic bacteria counts must be controlled to prevent biofilm formation. While absolute sterility isn’t required, bacterial counts should remain below 10,000 CFU/mL through continuous disinfection.
pH and alkalinity: Optimal pH ranges from 6.5-8.5 for most cooling systems. Alkalinity between 50-200 mg/L as CaCO₃ provides buffering capacity while limiting scale potential.
Tomi Engdahl says:
https://www.ramboll.com/data-centers-mission-critical-facilities/early-stage-works
Your early-stage guide to data centre site readiness
Early-stage works play a critical role in shaping successful data centres. Explore how infrastructure access, grid capacity, environmental considerations, community engagement, and planning and regulatory requirements come together to define what a site can realistically deliver.
Tomi Engdahl says:
In agricultural sprinkler irrigation, 60% to 85% of the water reaches the ground and crop root zone, while 15% to 40% is lost to evaporation and atmospheric drift.
Once water lands on the field (100% of the water that actually hits the ground), its journey splits into three main pathways: crop intake, soil surface evaporation, and groundwater recharge.
50% – 70% Taken up by roots, pulled through the crop, and released into the air through leaves (transpiration). This drives crop growth.
15-30% Direct evaporation from hot soil, puddle surfaces, and wet plant leaves before roots can absorb it. 10-20% Drainage past the root zone (percolation) into deep soil layers and underlying aquifers, or surface runoff to streams.
Tomi Engdahl says:
It’s not he amount of people, but the rate that farmers have been pumping water out of it over the last 60 years…
Tomi Engdahl says:
Nearly two-thirds of new U.S. data centers are built in high water-stress areas (such as Northern Virginia, Phoenix/Arizona, and West Texas). In small desert or High Plains towns, a single hyperscale AI data center can suddenly demand 20% to 30% of the entire municipality’s treated water budget.
Tomi Engdahl says:
Perspective: In the U.S., growing alfalfa and hay alone to feed livestock consumes thousands of times more water annually than every data center in the country combined claims Google Gemini. It also says
Data centers in the United States account for roughly 0.14% to 0.3% of total national freshwater withdrawals.
While the math behind those two numbers checks out, they measure two completely different things: absolute volume versus local infrastructure impact.
Tomi Engdahl says:
Modern liquid-cooled data centers generally use a two-loop design:
Primary (Facility) Loop: Runs between the outdoor cooling equipment (chillers, dry coolers) and the building. This uses standard industrial water or glycol blends.
Secondary (Rack/Chip) Loop: Circulates liquid through a Cooling Distribution Unit (CDU) directly to cold plates mounted on CPUs and GPUs. This is where DI water is used.
Tomi Engdahl says:
Modern liquid-cooled data centers generally use a two-loop design:
Primary (Facility) Loop: Runs between the outdoor cooling equipment (chillers, dry coolers) and the building. This uses standard industrial water or glycol blends.
Secondary (Rack/Chip) Loop: Circulates liquid through a Cooling Distribution Unit (CDU) directly to cold plates mounted on CPUs and GPUs. This is where DI water is used.
Closed-loop water cooling trades high water consumption for low-volume chemical stewardship. Threating the chemical mixture as a regulated industrial fluid rather than disposable wastewater. Flushing, or eventual decommissioning, strict protocols govern disposal: On-Site Deionization & Neutralization or Industrial Waste Hauling. Municipal industrial discharge permits restrict what can enter local sewers. For water based closed loop system chemicals are Antifreeze / Glycol, biocides and Corrosion Inhibitors.
Direct-to-chip or total liquid immersion cooling generally use one of two main categories of fluid in rack loop: Synthetic Oils / Hydrocarbons (Most Common) and Fluorinated Fluids / PFAS (Phase-Out).
In modern data centers, spent dielectric fluid is treated as an asset or regulated industrial waste. In the US, Europe, and many international jurisdictions, industrial fluids fall under strict hazardous/chemical waste tracking rules. Fluids that cannot be recycled are sent to licensed industrial incineration facilities that burn them at ultra-high temperatures.
Corporate Audit Exposure: Hyperscalers (Google, Microsoft, AWS, Meta) operate under strict environmental compliance audits. A single instance of illegal dumping resulting in groundwater contamination would trigger massive regulatory fines, EPA Superfund liability, and severe public backlash.
Tomi Engdahl says:
The Trump administration announced that it is exempting power plants that only serve data centers and do not connect to the broader grid from pollution limits that seek to prevent acid rain. https://thehill.com/policy/energy-environment/5992657-epa-data-center-power-plants-acid-rain/