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.
1,265 Comments
Tomi Engdahl says:
no water at all or very little water used in Finland for data center.
Finland: 0.00 to 0.05 L/kWh
USA: 1.80 L/kWh (US Dept. of Energy avg), hot regions 2.00 to 9.00 L/kWh
Tomi Engdahl says:
Finland heating:
Google Hamina data center makes 80% of heat for city if Hamina district heating network.
There are several data centers heating Espoo district heating network already. New Microsoft data center will heat 40% of Espoo when fully in use.
Many other cities have data centers that supply heat. Legislation demand waste heat reuse needs to be considered.
Tomi Engdahl says:
I rather see mostly CO2 electrical power running data center and waste heat heating my home than burning oil or wood for the same heating (through district heating). I lived through all options. Seen how data data center heat reduced the need to burn wood to only coldest winter days.
Tomi Engdahl says:
https://www.facebook.com/share/p/1CdypwMbgq/
Microsoft Has Signed a Deal to Buy Power From Three Mile Island — the Most Famous Nuclear Site in America
Three Mile Island Unit 1 — the reactor that remained operational after Unit 2′s partial meltdown in 1979 — was shut down by Exelon in 2019 due to low wholesale electricity prices. In September 2023, Constellation Energy announced it would restart the plant and signed a 20-year power purchase agreement with Microsoft to supply nuclear electricity for the tech giant’s growing data centre operations in Pennsylvania.
The deal has profound symbolic significance. Three Mile Island was the accident that defined American nuclear scepticism for 40 years. Restarting it — renamed Crane Clean Energy Center — as a clean power source for artificial intelligence infrastructure is the most visible symbol yet of nuclear energy’s rehabilitation in American public discourse.
The economics driving the restart are straightforward: Microsoft needs enormous quantities of 24/7 carbon-free electricity for data centres that AI workloads are making exponentially more power-hungry, and nuclear is the only scalable technology that delivers clean power at all hours regardless of weather. The 20-year contract provides the revenue certainty that made restarting Unit 1′s mothballed systems financially viable despite the significant capital required to return the plant to operational condition.
The broader implication for the US nuclear industry is substantial. If a mothballed nuclear plant can be economically restarted on the strength of tech sector demand, dozens of other shut or partially operational US nuclear facilities become candidates for similar arrangements. Holtec International has proposed restarting the Palisades plant in Michigan. Entergy’s Indian Point Unit 3 closure in New York is being revisited in political discussions.
America’s nuclear past is becoming its AI-powered future.
Source: Constellation Energy, Crane Clean Energy Center Restart Announcement 2023; Microsoft, Clean Energy Procurement Strategy 2023
Tomi Engdahl says:
https://www.facebook.com/share/p/1C6UnkoMVp/
The United States has deployed its first enhanced geothermal system specifically designed to power a hyperscale AI data centre, with Google signing a 20-year power purchase agreement for 115 megawatts from Fervo Energy’s Cape Station project in Utah.
Fervo uses horizontal drilling and hydraulic fracturing techniques borrowed from the oil and gas industry to create artificial reservoirs in hot dry rock. Water is injected into the fractured rock, absorbs heat, and is produced through a second well to drive turbines at the surface. Unlike solar and wind, geothermal provides constant baseload power, which AI data centres require to run around the clock without interruption. Google’s data centre electricity demand has grown faster than its ability to procure renewable energy from intermittent sources.
The Utah project will begin delivering power in 2026 and reach full capacity by 2028. The US Department of Energy estimates that enhanced geothermal could provide 90 gigawatts of firm capacity by 2050, up from about four gigawatts today. The DOE has awarded more than $99 million to 21 geothermal projects and announced a $171.5 million funding opportunity for next-generation geothermal field-scale tests and exploration drilling. The Bureau of Land Management will hold a competitive geothermal lease sale in Idaho offering 35 parcels totalling nearly 113,330 acres. Source: Fervo Energy, 2026.
Tomi Engdahl says:
https://www.facebook.com/share/p/1CsyjYzhaJ/
AI data center developers are increasingly pursuing “Bring Your Own Power” (BYOP) strategies, where they build their own power plants. These BYOP plants which are most often gas, frequently evade regulations and harm communities and the climate.
These massive BYOP gas projects will worsen climate change and harm the health of nearby communities, but it’s not too late to act. This is a critical moment to protect the climate, public health, and community self-determination.
Tomi Engdahl says:
https://www.facebook.com/share/p/1MPTskqvNo/
Finland is showing why cold climates don’t automatically rule out solar power.
Finland receives much less annual sunlight than southern European countries, but solar power has still been growing rapidly. Modern panels can generate electricity even in cold and snowy conditions, and snow can reflect additional sunlight toward the panels.
Bifacial solar panels are particularly interesting because they can capture light from both sides. When sunlight reflects off bright snow, the rear side of a bifacial panel can receive additional irradiance.
Cold temperatures can also improve photovoltaic efficiency, although winter’s short daylight hours and snow coverage remain important limitations.
Finland’s experience shows that solar energy isn’t limited to hot, sunny countries. The economics depend on electricity prices, system costs, panel orientation, snow management and local conditions.
Snow may look like a problem—but under the right conditions, it can also become part of the solar equation.
#SolarEnergy #Finland #BifacialSolar #RenewableEnergy
Tomi Engdahl says:
Yeah, water for wildlife but what about water for humans? Do we know how much water is wasted by the Mgm grand where they have the water sprinklers or irrigation? • Lawn Irrigation (U.S.): Consumes over 2 trillion gallons of water every single year to keep residential and commercial grass green. • Golf Course Irrigation (U.S.): Uses more than 500 billion gallons of water annually, easily outpacing direct data center cooling demands. I wonder how many people are willing to give up their green lawn or their green golf courses to save a little more water
Tomi Engdahl says:
Data center don’t use much water, like the cooling system in your computer. It gets recycled.
Tomi Engdahl says:
Because as of today I researched, Evaporative Cooling, the most common form of generative LLM data center cooling, uses evaporation towers by allowing water to evaporate off of the heated components, cooling them down, turning the water into vapour that (with the highest chance) turns into clouds that go into the sea, displacing it and making it no longer immediately available for the local public.
Tomi Engdahl says:
https://www.facebook.com/share/p/1DkKucnuRw/
The AI boom is not free. And the question of who pays for it is now being fought in court.
A Bloomberg News analysis of wholesale electricity prices found that in areas near significant data center activity, costs are as much as 267% higher for a single month than they were five years ago. That is the wholesale price utilities pay producers. It is not the final rate on your bill, but it is the cost that gets passed down to households and businesses on the same grid.
On the PJM grid—the largest in the U.S., serving 67 million people across 13 states—the numbers are stark. PJM’s independent market monitor, Monitoring Analytics, estimated that data center demand has added nearly $30 billion to wholesale electricity costs since 2024. That is almost half of all capacity costs charged during that period.
In March 2026, Microsoft, Amazon, Google, Meta, xAI, Oracle, and OpenAI signed a “Ratepayer Protection Pledge” at the White House, voluntarily agreeing not to raise people’s power bills. The pledge asked tech entities to shield everyday consumers by covering the cost of the new power plants and grid infrastructure built to serve them.
But in September 2026, Microsoft filed a formal appeal with the Virginia Supreme Court to challenge a State Corporation Commission order requiring data center operators to pay upfront for transmission infrastructure. Microsoft argues that Virginia’s localized prepayment rule conflicts with broader federal standardization efforts. Critics say fighting the binding state-level mechanism effectively undermines the intent of the public pledge.
The takeaway is not that AI is bad. It is that someone has to pay for the infrastructure. And the fight over who that is has just begun.
SOURCES: Bloomberg | WRAL | PolitiFact | Monitoring Analytics | The Verge | Virginia State Corporation Commission | Energy Central | Cleveland.com | U.S. House Energy and Commerce Committee
Tomi Engdahl says:
https://www.facebook.com/share/p/1FRzSWFk5G/
An incident involving emergency diesel generators at a data center in Virginia has drawn attention to a frequently overlooked part of the digital economy: the machinery that keeps data flowing when the power supply is interrupted.
Data centers support cloud computing, online services, artificial intelligence, and countless everyday applications. To maintain continuous operations, many facilities use backup generators that can supply electricity during outages or other disruptions. These systems are important for reliability, but they can also create noise and exhaust emissions when operating.
For nearby residents, the impact depends on how many generators run, how long they operate, their distance from homes, and whether effective noise and emissions controls are in place. Even equipment designed for emergencies can become a community concern if testing or operation causes repeated disturbances.
The incident highlights why data-center planning must include more than computing capacity and electricity demand. Operators also need transparent emergency procedures, suitable equipment placement, environmental monitoring, and communication with neighboring communities.
As artificial intelligence expands the demand for data infrastructure, responsible development will require balancing digital reliability with public health, environmental protection, and the quality of life of people living nearby.
Tomi Engdahl says:
https://www.facebook.com/share/p/1DhqSDZSiu/
Data centers are designed to operate around the clock, and the equipment inside them releases heat as it consumes electricity. Researchers are examining whether the concentration of these facilities can influence temperatures in surrounding areas.
A study associated with researchers at the University of Cambridge has drawn attention to the possibility of localized warming around data-center infrastructure. Reported temperature effects can vary according to the facilities being studied, the surrounding land, weather conditions, and the methods used to measure heat.
The basic mechanism is straightforward. Servers generate heat, and cooling systems transfer that heat away from equipment. Depending on the facility’s design, some of that heat is released into the surrounding air. Large buildings, paved surfaces, and supporting infrastructure can also influence local temperature patterns.
However, claims about a specific temperature increase or a warming distance of several miles should be interpreted in the context of the study’s methods and geographic scope. A modeled or observed effect at one location does not automatically apply to every data center.
As AI computing expands, planners need to consider more than electricity demand. Cooling efficiency, waste-heat reuse, building design, local climate, and the location of new facilities all matter.
Data centers are essential to many digital services, but their physical footprint deserves careful study. Better measurements can help communities understand where localized impacts occur and how engineering choices might reduce them.
Tomi Engdahl says:
https://www.foodandwaterwatch.org/2026/09/25/data-centers-and-carbon-capture-a-dangerous-alliance/?ms=some_fb_10062026_NAT-FWW-c3-data-centers-ccs-web&oms=some_fb_10062026_NAT-FWW-c3-data-centers-ccs-web&fbclid=IwdGRjcAU5UPNjbGNrBTlQ7GV4dG4DYWVtAjExAHBkb2YFc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeAjZLYs09LC3h0AkKLQVKFtC7dl6TIqQRrpJZ_gm21gSWdD5KfRDd2wncuwQ_aem_gSp0WXNWBAzDxTAc3Pd40Q
Tomi Engdahl says:
https://www.facebook.com/share/p/1QKHgsyt5w/
Data centers for AI burn nearly half of America’s new power demand.
The International Energy Agency expects data centers to make up nearly half of all growth in US electricity demand through 2030. Nearly half of the growth in US power demand through 2030 comes from data centers, many of them built for AI.
In the US, coal and gas still make up just under 60% of the power supply, MIT Technology Review reports. Google says a typical text prompt to its Gemini AI uses only 0.24 watt-hours, about nine seconds of TV.
Should new power plants serve AI data centers first, or should homes and businesses come first?
Sources: MIT Technology Review, International Energy Agency
Tomi Engdahl says:
https://www.nbcwashington.com/news/local/northern-virginia/data-center-light-noise-pollution-reston-residents/4165305/?_osource=SocialFlowFB_DCBrand&fbclid=IwdGRjcAU5pEZjbGNrBTmkO2V4dG4DYWVtAjExAHBkb2YFc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEeQawZhpWsUd_AwYFFSp-4WAckdBlSZc6z–7z3oFTFmrpJw0kObX9CowRF_U_aem_n5OiLfHY-3T1YdBb_ZOGPQ
Tomi Engdahl says:
https://www.facebook.com/share/p/19YoLPexfu/
Georgia Power Reverses Retirement Plans For The Massive Plant Scherer Coal Station
One of the largest coal-fired power stations in the United States is expected to continue operating after Georgia Power reversed its earlier plans to retire Plant Scherer in Georgia. The decision reflects rising electricity demand associated with data centers, artificial intelligence infrastructure, and industrial development.
Located near Juliette, Georgia, Plant Scherer was built in stages, with its first unit entering service in 1982. The facility originally had approximately 3,720 megawatts of generating capacity. Following the retirement of Unit 4 in 2022, three units remained in operation.
The plant had previously been scheduled for retirement in December 2028 under Georgia Power’s earlier resource planning. Updated plans subsequently proposed maintaining the remaining generating capacity to help meet projected electricity requirements. The revised approach reportedly received approval from the Georgia Public Service Commission, although the precise operating timeline depends on current regulatory and planning documents.
Georgia Power argues that retaining existing generating assets can help address growing demand while avoiding the immediate expense of replacing their capacity. The company has also highlighted investments in pollution-control equipment designed to reduce emissions.
However, the decision remains controversial. Residents and environmental advocates have raised concerns about coal ash, groundwater, air pollution, and the long-term consequences of continued coal combustion.
The debate illustrates a central challenge in the energy transition: balancing reliable electricity supplies and rising demand against environmental protection and the shift toward cleaner energy sources.
Source: Georgia Power; Georgia Public Service Commission; Plant Scherer resource-planning documents.
Tomi Engdahl says:
https://www.facebook.com/share/1ZUjuQNyPh/
In 2025, wind turbines generated about 19 TWh of electricity in Denmark, supplying roughly 60% of the country’s domestic electricity production. Wind output was about 1.5 TWh lower than in 2024 because of weaker wind conditions.
Despite the drop, wind remains Denmark’s biggest source of electricity and a key part of its clean-energy strategy. The country is aiming to reach 100% clean electricity by 2030, relying heavily on wind power and other renewable sources.
Denmark’s experience shows how a country can integrate large amounts of variable renewable energy into its power system. However, low-wind periods still require electricity imports, grid interconnections and flexible generation or storage to maintain supply.
Source:
Scandinavia Standard
Space Daily
Tomi Engdahl says:
https://tieku.fi/teknologia/nyt-datakeskukset-lahtevat-avaruuteen-ja-teknologiajatit-ryhtyvat-uuteen-avaruuskilpailuun
Tomi Engdahl says:
https://www.birdsadvice.com/ai-data-centers-are-tied-to-a-staggering-amount-of-water-use-across-7-states/?utm_source=A34&utm_id=014&fbclid=Iwb21leAUtZb1jbGNrBS1lsmV4dG4DYWVtAjExAHBkb2YFc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEe-erkXT3oGIag-sa1vvrcyKuIp0HsEAZClPI8FY_OhsMeRWItnjNxSY544HE_aem_XLqqLQ1GmYtaj7WJbaqwKA
Tomi Engdahl says:
$10,000 to let them build a data center near you. Would you have taken it?
Source: https://go.ufdmedia.com/ZDpQ5l
Tomi Engdahl says:
https://www.facebook.com/share/p/1AsMD9iZh5/
Artificial intelligence is changing how people work, create, and communicate, but the infrastructure behind it comes with environmental costs. Data centers require electricity, cooling systems, water in some locations, and substantial resources to manufacture the equipment that keeps them running.
That doesn’t mean every AI-generated image carries the same environmental impact, or that technology cannot become more efficient. It does mean the conversation about AI should include energy demand, water use, emissions, and transparency. As digital technology grows, protecting the environment deserves a place in decisions about how it develops.
Tomi Engdahl says:
https://www.facebook.com/share/p/19oSHMVYMW/
A growing health crisis in eastern Oregon is raising alarm, with residents linking nearby Amazon data centers to rare cancers, pregnancy complications, and contaminated groundwater.
In Morrow County, the rapid expansion of Amazon Web Services data centers has sparked serious public health concerns. Investigations found that the facilities’ cooling systems draw millions of gallons of local groundwater daily. As that water evaporates to cool thousands of servers, existing agricultural contaminants, especially toxic nitrates, become highly concentrated before being discharged back into local aquifers. Well tests in surrounding communities have recorded nitrate levels up to seven times above federal safety thresholds.
Residents and healthcare advocates report a troubling rise in rare soft-tissue sarcomas, non-Hodgkin lymphoma, and pregnancy complications. Amazon attributes the contamination to decades of regional agricultural runoff, while environmental experts argue the intense water cycling has supercharged pollutant levels in vulnerable rural communities. As AI and cloud computing demand ever more energy and water, the case highlights the urgent need for stricter environmental oversight of tech infrastructure.
#DataCenters #Oregon #WaterCrisis #EnvironmentalHealth
Tomi Engdahl says:
Väitteessä on totuuden siemen, mutta se on yleistävä. Kaikki sisäisen kierron jäähdytysjärjestelmät eivät perustu PFAS-yhdisteisiin.
Suurin osa moderneista sisäisen kierron järjestelmistä (kuten tavallinen Direct-to-Chip -nestejäähdytys) käyttää tavallista vesi-glykoliseosta tai biohajoavia synteettisiä esteriöljyjä, jotka ovat täysin PFAS-vapaita. Lisäksi teollisuus siirtyy parhaillaan kovaa vauhtia pois kaikista PFAS-pohjaisista nesteistä tiukentuvien ympäristölakien vuoksi.
Tomi Engdahl says:
https://www.facebook.com/share/p/1CEqfn3Vz5/
Sam Altman Says 38,000 ChatGPT Queries Use as Much Water as One Almond
OpenAI CEO Sam Altman said that about 38,000 ChatGPT queries use roughly as much water as growing one California almond. He also acknowledged that the figure was recalled from memory and may not be exact.
A 2019 study estimated that one California almond requires about 3.56 liters of water, which would put the comparison closer to 11,000 ChatGPT queries, or roughly one-third of Altman’s estimate. The difference shows why AI water-use comparisons should be treated carefully.
The comparison also sparked humorous criticism online, with people noting that an almond provides food while thousands of AI queries provide no calories. Behind the debate is a serious issue: AI data centers consume water for cooling, while the industry still provides limited public data on the water used by individual AI services.
Tomi Engdahl says:
Eräät kemikaaliteollisuuden toimijat, näkyvimpänä Chemours, lobbaavat voimakkaasti sen puolesta, että PFAS-yhdisteet (fluoropolymeerit ja tietyt F-kaasut) ovat välttämättömiä (essential use) paitsi vihreän siirtymän teknologioille (tuuli- ja aurinkovoima, vetytalous, sähköautojen akut), myös nyt rajussa kasvussa olevalle tekoäly- ja datakeskusinfrastruktuurille.
Uusi markkina tekoälylle: Vaikka turvallisempia vaihtoehtoja on olemassa (kuten PFAS-vapaat vesipohjaiset järjestelmät tai hiilivedyt), kemikaalijätit markkinoivat aggressiivisesti fluorattuja nesteitä erityisesti kahden faasin (two-phase) upotusjäähdytykseen, jossa nesteet kiehuvat ja höyrystyvät suljetuissa altaissa äärimmäisen tehokkaasti.
Kaksifaasinen PFAS-jäähdytys oli pitkään teknisesti lupaus äärimmäisestä tehokkuudesta, mutta sen käyttö on ajautunut kriisiin, koska
EU:n ja US EPA:n tiukentuvat PFAS-kiellot ovat saaneet suuret hyperskaalaajat (Microsoft, Intel, Google) luopumaan tai jäädyttämään laajamittaiset tätä tekniikkaa käyttävät hankkeensa.
Teollisuus on siirtänyt katseensa nopeasti täysin PFAS-vapaaseen yksifaasiseen upotusjäähdytykseen (Single-Phase Immersion) sekä muihin suoranestejäähdytysratkaisuihin.
Yksifaasisessa järjestelmässä neste ei kiehu, vaan se pysyy aina nestemäisenä ja sitä kierrätetään pumppujen ja lämmönvaihtimien läpi.
3M – alan suurin valmistaja – lopetti kaikkien PFAS-kemikaaliensa tuotannon vuoden 2025 loppuun mennessä miljardiluokan ympäristösyytteiden vuoksi. Tämä halvaannutti markkinat.
Vaikka Chemours yrittää lanseerata uusia fluorattuja HFO-nesteitä (kuten Opteon 2P50) datakeskuksille, moni konesalioperaattori pitää niitä liian suurena lainsäädäntö- ja maine-eriskinä.
Tuotannon kasvuriskit: Ympäristöjärjestöjen (kuten ChemSecin) ja tutkijoiden tuoreet raportit vahvistavat esittämäsi huolen: tekoälydatakeskusten ja vihreän teknologian räjähdysmäinen kasvu uhkaa jopa moninkertaistaa PFAS-kemikaalit tuotannon, ellei lainsäädäntö pysäytä kehitystä.
https://chemsec.org/data-centres-dont-need-pfas-cooling-systems/?hl=en-US#:~:text=One%20more%20detail%3A%20when%20the%20liquids%20in,most%20of%20all%20it%20has%20proven%20reliable.
https://www.nrdc.org/media/chemours-and-data-centers-love-story?hl=en-US#:~:text=For%20example%2C%20evaporative%20cooling%20consumes%20large%20quantities,low%20doses%20(in%20the%20parts%20per%20trillion).
https://thecoolingreport.com/intel/pfas-two-phase-immersion-cooling-crisis.html?hl=en-US#:~:text=See%20plans.%20On%20December%2020%2C%202022%2C%203M,7100%2C%20Novec%20649%2C%20Fluorinert%20FC%2D72%2C%20are%20gone.
https://www.nrdc.org/media/chemours-and-data-centers-love-story?hl=en-US#:~:text=In%20addition%20to%20being%20highly%20persistent%2C%20PFAS,shorter%2Dchain%20PFAS%20that%20persist%20in%20the%20environment.
Tomi Engdahl says:
Missä PFAS-kemikaaleja käytetään datakeskuksissa?
PFAS-yhdisteet ovat löytäneet tiensä datakeskuksiin neljässä pääkohteessa:
1. Lämpöpumput ja perinteiset ilmastointilaitteet (Kylmäaineet/F-kaasut):
Suurin osa maailman datakeskuksista käyttää edelleen perinteisiä kompressoripohjaisia ilmastointi- ja lauhdutinjärjestelmiä. Monet näissä käytettävistä synteettisistä kylmäaineista (kuten fluoratut HFC- ja HFO-kaasut, esim. R-134a tai R-1234ze) luokitellaan kemiallisesti PFAS-yhdisteiksi. Tämä on korkein ja yleisin PFAS-käyttökohde alalla tällä hetkellä.
2. Kaasumaissammutusjärjestelmät (Fire Suppression):
Serverisaleissa ei voida käyttää vettä tulipalon syttyessä. Palonsammutusjärjestelmissä käytettävät “puhdaskaasut” (clean agents, kuten Novec 1230 eli FK-5-1-12) ovat synteettisiä fluorattuja kemikaaleja (PFAS), joita pidetään palvelinsalien standardina.
3. Komponentit ja kaapelointi:
Palvelinten emolevyillä, suurnopeuskaapeleiden eristeissä (PTFE/Teflon) ja tiivisteissä käytetään PFAS-pohjaisia fluoropolymeerejä niiden erinomaisen lämmön- ja sähköneristyskyvyn ansiosta.
4. Erikoisnesteet (Kaksifaasinen upotusjäähdytys):
Huipputehokkaissa tekoälydatakeskuksissa kokeillut suoraan kiehuvat dielektriset nesteet (kuten 3M:n Novec-nesteet) olivat puhtaita PFAS-kemikaaleja. Tämä edustaa kuitenkin vain pientä osaa koko maailman datakeskuskapasiteetista.
Tomi Engdahl says:
Mikä on tilanne tällä hetkellä ja mihin suuntaan ala menee?
Käyttö on murroksessa tiukentuvan lainsäädännön ja kemikaalivalmistajien ratkaisujen vuoksi:
Jäähdytysnesteet vaihtuvat:
Erikoisnesteitä vaativassa upotusjäähdytyksessä ala on siirtymässä PFAS-vapaaseen yksifaasiseen upotukseen (synteettiset esteriöljy- ja hiilivety-yhdisteet) tai tavalliseen vesi-glykolipohjaiseen suorasirunestejäähdytykseen (Direct-to-Chip).
Kylmäaineiden sääntely:
Euroopan unionin F-kaasuasetus ja Yhdysvaltain EPA-sääntely ajavat datakeskuksia vaiheittain eroon korkean ilmastovaikutuksen ja PFAS-luokituksen saaneista kylmäkaasuista, ja tilalle tuodaan luonnollisia kylmäaineita (kuten propaania R290 tai CO₂:ta).
Korvaamattomat kohteet:
Elektroniikkakomponenttien PTFE-eristeille ja piirikorttien suojauksille ei vielä toistaiseksi ole yhtä suorituskykyisiä vaihtoehtoja, minkä vuoksi kemikaaliteollisuus lobbaa näille tuotteille erivapauksia.
Tomi Engdahl says:
Perinteisissä ilmastointilaitteissa, lämpöpumpuissa ja kylmäkoneissa PFAS-yhdisteitä on käytetty synteettisinä kylmäaineina (kuten monet HFC- ja HFO-kaasut, esim. R-134a tai R-1234ze).
Kun sääntely ja PFAS-kiellot tiukentuvat, teollisuus siirtyy korvaamaan näitä synteettisiä kaasuja luonnollisilla kylmäaineilla (Natural Refrigerants), jotka ovat täysin PFAS-vapaita. Tärkeimmät korvaajat ovat:
1. Propaani ja muut hiilivedyt (Hydrocarbons – esim. R290)
2. Hiilidioksidijärjestelmät (CO2 R744)
3. Ammoniakki (NH3 / R717)
4. Vesi (H2O / R718) & Ilma (R729)
Modernissa LVI- ja datakeskustekniikassa PFAS-kaasujen paras korvaaja on paluu luonnollisiin aineisiin (erityisesti propaaniin R290 ja CO2:een). Ne eivät kerry ikuisuuskemikaaleina luontoon, eivät vahingoita otsonikerrosta ja voivat toimia jopa energiatehokkaammin kuin vanhat synteettiset PFAS-kaasut.
Tomi Engdahl says:
“These facilities could lock America into decades of continued fossil fuel dependence.” https://trib.al/LJAShQv
Spew Baby, Spew
The AI Data Center Boom Is Causing a Gigantic Spike in Smog-Spewing Gas Power Plants
“These facilities could lock America into decades of continued fossil fuel dependence.”
https://futurism.com/future-society/ai-data-center-boom-smog-spewing-gas-power?fbclid=IwdGRjcAU7M9VjbGNrBTszxmV4dG4DYWVtAjExAHBkb2YFc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEe-wy9jGDfVebHBhDbK09GwJ_7mAPGRODR4RQJfQ0zlmzUTSd54wck24w9gzU_aem_WdDQ6iaofBEf6cLPwqctsw
The United States runs on gas powered generators.
As of 2022, the US had nearly 1,000 gas-fired plants, which was already more than the next nine countries combined. Under returning president Donald Trump, the country’s embrace of high-carbon power has only accelerated, as energy developers abandon environmentally-friendly projects like off-shore wind farms in favor of more noxious gas- and coal-powered facilities.
As if our rapidly warming world wasn’t being choked by enough smog, the AI boom is now adding even more fuel to the proverbial fire. According to a new analysis by the Environment America Research and Policy Center, first flagged by Electrek, the buildout of more AI infrastructure will come with a 44 percent increase in gas-fired generator capacity over the next four years.
That’s a stunning rise of nearly 20 gigawatts by 2030, compared to the country’s current gas generator capacity of 41.8 gigawatts, a milestone which took centuries of development — and untold tons of carbon exhaust — to reach.
As Electrek notes, the planned expansion of gas capacity exceeds the combined output of every single US state except for Texas, California, and Florida. And as more noisome gas facilities appear on the American horizon, planned green energy installations are being axed — with a 34 percent decrease in planned wind farm capacity since December of 2025.
“The data is clear: surging electricity demand to power data centers is putting our nation’s transition to clean energy at risk,” chief operating officer for The Public Interest Network Andre Delattre told Electrek. “Given the lifespans of new gas plants, these facilities could lock America into decades of continued fossil fuel dependence. Any potential benefits of data centers won’t mean anything if we destroy the planet in the process.”
Tomi Engdahl says:
https://www.facebook.com/share/r/1HzPnuXCNC/
Virginia’s Department of Environmental Quality released a major groundwater assessment warning that the Potomac Aquifer and surrounding eastern groundwater system cannot reliably support additional large industrial withdrawals without risking water availability. The findings could affect water-intensive data-center projects across the region.
Important Details:
Potomac Aquifer — A major groundwater source serving parts of eastern Virginia.
State scientists warned that new major industrial water withdrawals could worsen groundwater depletion.
Falling groundwater levels can contribute to land subsidence and saltwater intrusion.
Water-intensive data centers using evaporative cooling could face serious supply constraints.
Amazon changed its water-supply approach for a proposed King George County campus after concerns about groundwater availability.
Virginia’s findings underline a growing infrastructure challenge: AI data centers need not just electricity, but dependable water supplies in regions already facing environmental limits.
#Virginia #DataCenters #AI #WaterCrisis #Environment
Tomi Engdahl says:
“Not a single drop has come through our taps for the last three months,” a local tells the BBC. Read more: https://bbc.in/4zp8uET
Tomi Engdahl says:
https://www.facebook.com/share/p/1FYAMEFc5c/
Texas oilfields generate around 8 billion gallons of wastewater daily. Researchers think some of it could help cool America’s thirsty data centers.
Texas oilfield wastewater could help cool data centers without draining freshwater
Texas researchers propose treating oilfield wastewater to cool data centers and reduce their dependence on freshwater.
https://interestingengineering.com/energy/texas-oilfield-wastewater-data-center-cooling?fbclid=IwdGRjcAU8H3djbGNrBTwfZGV4dG4DYWVtAjExAHBkb2YFc3J0YwZhcHBfaWQMMzUwNjg1NTMxNzI4AAEe0Cz9jkl4fr0XGIjhOgcfr12KRh4dzvCbV-2oe_pv-RWP9J4pTuuXP-2Albo_aem_HfmTD1fF833suRZQQaIfYQ
Can waste water from the oil and energy industry quench the thirst of data centers? Researchers at The University of Texas in Austin are betting it can.
They suggest that established desalination technologies could allow data centers to use salty or contaminated water to meet their cooling needs, reducing their reliance on freshwater.
Although there are various ways to separate salt from water, researchers said reverse osmosis, which uses pressure to push water through a semipermeable membrane, is more energy-efficient. However, the technology has one drawback: it struggles to treat water saltier than seawater.
Heating water into vapor and then condensing it back into freshwater is another viable solution, but this method requires more energy.
Turning wastewater into coolant
According to the Environmental and Energy Study Institute, large data centers can consume up to 5 million gallons of water per day. The waste water generated by certain oil fields in Texas is over thousand fold higher than this amount.
Data centers can become more efficient in their water usage if they reduce stress on local water resources without drastically increasing costs or overall energy demand, according to a recently published study in Water Research.
“We’re not waiting for a futuristic technology to come along and solve this problem for us,”
Bahudur noted that though the desalinations methods are available but they incur a cost. These costs, however, will be a small fraction of a data center’s total energy footprint when compared even with the most energy-intensive water treatment.
Texas oilfields offer alternatives
He proposed turning attention to the Permian Basin. This geological region in West Texas generates approximately 8 billion gallons of produced water per day during oil and gas extraction.
The researcher said this water could be put to better use instead of being transported elsewhere for disposal, where it could pose long-term risks.
“The Permian (Basin) puts out 1,200 Olympic-sized swimming pools per day, and right now this water is transported over long distances and injected underground. This is problematic in many ways including the risk of seismic activity,” Bahadur said. “We’re also running out of space to bury that water, by as early as 2030. We need to find a way to use it, and cleaning it up for data centers is a great way to do that.”
The research finding becomes significant given that only less than 4 percent of Earth’s water is freshwater, and only a fraction of it, just 0.5 percent, is accessible and safe for human consumption. Additionally, by 2028, water consumption by data centers is expected to reach approximately 33 billion gallons, filling nearly 50,000 Olympic-sized pools, according to Lawrence Berkeley National Laboratory
Tomi Engdahl says:
https://climate.mit.edu/sites/default/files/2025-02/lbnl-2024-united-states-data-center-energy-usage-report.pdf?fbclid=IwSU9FWAU8ICJleHRuA2FlbQIxMABwZG9mBXNydGMGYXBwX2lkDDM1MDY4NTUzMTcyOAABHvS0I2hOOvaSkvwzE9OpevemFyA7H2h2J5fV-inxJGnXEfYeMWFJDLByRNNQ_aem_AkwVcsPylF3pg6LLmvysmg
Tomi Engdahl says:
https://www.newsweek.com/map-shows-the-13-states-under-emergency-order-to-prevent-blackouts-12560664?utm_medium=Social&utm_source=Facebook#Echobox=1791908559
Tomi Engdahl says:
https://www.facebook.com/share/r/1UrZi3RUDR/
Ireland’s data centers now consume 23% of the country’s electricity.
Data centers are buildings full of computer servers that store data and run online services, including AI tools, for tech companies.
That is more electricity than every urban home in Ireland combined, which used 18%, according to Ireland’s Central Statistics Office.
Google, Meta, Amazon, Microsoft and TikTok all run data centers in Ireland, and 97% of the country’s data centers sit around Dublin.
In 2021, Ireland’s energy regulator restricted new data center grid connections, and grid operator EirGrid signaled none would be approved around Dublin before 2028.
Since December 2025, new data centers must bring their own power generation or storage, on site or nearby, to connect to Ireland’s grid.
Ditch big tech and use the Conscious Spend app to find ethical alternatives.
Tomi Engdahl says:
Reading through these comments, one thing is obvious: most people don’t actually understand how modern data center thermal management works.
Zero Evaporation: Modern facilities use closed-loop cooling. Unlike old open cooling towers that constantly evaporate water into the air, a closed loop operates like a giant sealed radiator. Once filled, the fluid recirculates continuously for years without daily water draws.
It’s Not Plain Tap Water: Raw municipal water cannot be piped directly into server loops, it would ruin the heat exchangers. The fluid is hyper-purified, treated, and mixed with glycol and anti-corrosion inhibitors. When available, facilities use reclaimed “purple” water or industrial gray water, not local drinking water.
Some Actual Math: A typical closed-loop system holds around 100,000 to 200,000 gallons of actual water for a large facility hall. That water stays sealed in the system for up to 10+ years. Spread over a decade, that equals roughly 30 to 55 gallons of water per day, less than what a single average household uses to run a few loads of laundry.
Stop panicking over headline numbers you don’t understand. The engineering facts don’t back up the outrage.
Tomi Engdahl says:
That sentiment captures the core emotional and social friction driving local pushbacks against tech infrastructure.
While the local impact on a farming community is dramatic, the macro scale reveals a stark contrast in land consumption:
Agricultural Land: In the U.S. alone, crop cultivation and pastureland span over 900 million acres.
Data Center Land: All commercial and hyperscale data centers combined across the entire United States occupy an estimated 100,000 to 150,000 acres total—roughly the size of a single large county.
The issue isn’t that data centers are paving over the entire planet; it’s that hyper-concentration causes severe localized strain on rural towns, county zoning boards, and local electrical grids that were never designed to power the equivalent of a small city.
Interestingly, some modern engineers and agricultural planners are trying to flip this conflict upside down by using data center waste heat to grow food.
Greenhouse Heating: In colder climates (like Northern Europe), data center waste heat (40 to 60 C water) can be piped directly into commercial greenhouses or vertical farms. This slashes the heating costs for year-round tomato, cucumber, or leafy green production, especially during harsh winters.
While data centers take up a tiny fraction of global land compared to farming, their rapid, power-hungry expansion into rural and agricultural zones forces communities to ask hard questions about zoning, water rights, and whether local utility rates should subsidize tech growth at the expense of regional resources.
Tomi Engdahl says:
The blanket statement “they use too much water” hits on a real issue, but lumps two completely different ways of building data centers into one.
Whether a data center uses “too much water” or zero water at all depends entirely on the design chosen by the engineers and approved by local zoning boards.
Tomi Engdahl says:
It seems the more intelligent the computers are getting the less intelligent people are getting
Tomi Engdahl says:
No single entity or company unilaterally decided that data centers (DCs) must exist. Their necessity is the result of a 40-year structural shift in how human civilization stores information, coordinates supply chains, executes economic trade, and conducts communication.
Data centers were not forced on the world by a single entity; they evolved to solve a physical limit. As society transitioned from physical paper, local analog media, and localized corporate servers to real-time global compute, centralizing silicon in dedicated thermal and electrical buildings became the only thermodynamically and economically viable way to run the modern world.
Tomi Engdahl says:
https://news.microsoft.com/source/features/sustainability/project-natick-underwater-datacenter/
Tomi Engdahl says:
Better Dead Than Red Several major technology companies and specialized hyperscalers have officially modified their infrastructure blueprints to eliminate evaporative cooling in new site builds, shifting entirely to zero-water, closed-loop, or air-cooled designs.
The transition is driven by local water permitting pushback, corporate “Water Positive” ESG targets, and the high thermal density of modern AI chips (which suit direct-to-chip closed liquid loops far better than outdoor wet towers).
Major Companies Phasing Out Evaporative Cooling: Microsoft, Meta, Amazon, Oracle, xAI
While Google has committed to being Water Positive by 2030, it remains one of the largest users of evaporative cooling towers in legacy facilities. However, Google is actively shifting toward recycled/greywater sourcing (using treated sewage effluent instead of drinking tap water) and expanding closed-loop dry cooling in newly planned campuses.
Tomi Engdahl says:
There might not be better ways than data centers to handle all of the modern digital things.
Because quitting technology is not an option for most people, the option is to build data centers and software better. The way data centers have been built and powered traditionally in USA is far from best way for environment.
For better ways look at what happens in Finland and Sweden.
Also look at “green code” movement that tries to reduce the environmental damage caused by software running our modern services. Optimizing software can potentially save a lot.
Tomi Engdahl says:
Data centers don’t necessarily need fresh water. They need a way to remove the enormous amount of heat generated by their computer servers. Fresh water is often used because it is relatively inexpensive, readily available, and works well in cooling systems—but there are alternatives, including reclaimed wastewater, seawater, air cooling, and specialized cooling fluids.
The important distinction is that water is used to move heat, not because computers require it to operate.
Tomi Engdahl says:
Traditional data centers that just do network traffic are air cooled. But anything requiring a GPU is liquid cooled, because air cooling is not sufficient enough to cool thousands of GPUs. If you’ve been part of building data centers that don’t require liquid cooling, you are most likely building the network building.
Tomi Engdahl says:
Myth 1: Data centers are completely automated and don’t create local jobs.
The Myth: People often assume that once a massive data center is built, it runs entirely on its own with a handful of remote engineers, offering little to no economic benefit to the host community. Myth 2: Data centers deplete local water supplies and leave communities dry.
The Myth: It is frequently stated that all data centers consume massive quantities of municipal drinking water continuously through traditional cooling towers.Myth 3: Data centers automatically make residential electricity bills skyrocket.
The Myth: A common belief is that data centers place an uncompensated burden on the local grid, forcing everyday residents to shoulder higher energy costs. Myth 4: Data centers are dead-weight energy hogs that offer zero grid value.
The Myth: Data centers are viewed solely as massive power sinks that drain energy resources without giving anything back. Myth 5: Data centers are whisper-quiet and invisible neighbors.
The Myth: Because data centers look like standard commercial warehouses, people often assume they have little to no environmental presence outside their walls.
Tomi Engdahl says:
DATA Center can produce significant noise pollution, primarily from the cooling systems (fans, chillers, and air handling units) required to manage heat from servers, as well as backup diesel generators and on-site power turbines.
Noise Levels: Traditional data centers typically emit 55 to 85 decibels, but dense AI facilities and those with large backup generators can reach 96 to 105 decibels (comparable to a lawnmower or jet engine).
Primary Sources: The constant hum is generated by HVAC systems, air-cooled chillers, and server fans; diesel generators add intermittent, louder rumbling during tests or outages.
Health Impact: Residents often report low-frequency noise that is difficult to measure with standard decibel meters but causes sleep disturbances, headaches, and cardiovascular stress.
Variability: Noise levels depend heavily on the facility’s design, cooling technology (air vs. liquid), and proximity to residential areas, with some modern facilities using sound barriers or remote locations to mitigate impact.
A typical mid-sized data center uses about 12,500 gallons (approx. 47,300 liters) of water per hour. Large or AI-focused facilities can consume up to 208,000 gallons (approx. 787,000 liters) per hour, heavily depending on outside weather, cooling design, and server load.
CONSTANT humming masks the calls animals use to find mates, warn of predators, or defend territories.
Behavioral Changes: Sensitive species often abandon traditional feeding, nesting, or breeding grounds near noisy industrial sites.
Stress and Health: Chronic noise raises stress hormone levels in animals, leading to long-term physiological impacts similar to those seen in humans.
Nocturnal Disruption: Continuous sound combined with perimeter security lighting heavily disorients nocturnal animals like bats, owls, and local insects.
Tomi Engdahl says:
The European Union has transitioned from voluntary guidelines to mandatory, legally binding efficiency standards and public reporting framework for data centers. The EU regulates data center energy efficiency through three progressive regulatory mechanisms: Mandatory Reporting, EU Sustainability Ratings, and Eco-Design Requirements.
Under EED Article 11 & 12, EU Member States require data center operators to analyze the economic feasibility of waste heat reuse.
Under Article 12 of the recast Energy Efficiency Directive (Directive EU/2023/1791), all data centers operating within the EU with an installed IT power demand of 500 kW or more must report performance and sustainability metrics annually.
Key Metrics Collected (Delegated Regulation EU/2024/1364)
Data center operators must submit standardized data across key performance indicators (KPIs) to the European Union database:
Power Usage Effectiveness (PUE): Ratio of total facility energy consumed vs. energy consumed strictly by IT equipment.
Water Usage Effectiveness (WUE): Volume of water consumed on-site per kWh of IT energy.
Energy Reuse Factor (ERF): Percentage of facility waste heat recovered and utilized off-site (e.g., piped into municipal district heating grids or greenhouses).
Renewable Energy Factor (REF): Proportion of total energy sourced from renewable supplies (PPAs, on-site solar/wind, green tariffs).
IT Hardware Efficiency & Capacity: Active server workload ratings (SERT scores) and new storage device installs.
Tomi Engdahl says:
This claim is partially true, but oversimplified.
Data Centers do NOT use water. They use Fans, AV and sealed closed loop cooling with a non corrosive fluid. Worry about all the power they use instead.
Water Usage Depends on Design: Modern, next-gen builds (especially for high-density AI clusters) increasingly use closed-loop fluid systems and outdoor dry radiators that consume zero operational water on-site. However, about 50% of operating data centers globally still rely on evaporative cooling towers, which evaporate millions of gallons of municipal water daily to lower electricity draw.
Power Draw is the Larger Strain: The second part of the claim is spot on. Electrical power consumption—not water—is currently the single largest bottleneck for data centers.