By David Wojick
False fears of a flood of data centers is warping the electric power policy debate. Here are three little numbers that tell the story. Well, actually, it is two little real numbers and one huge fantasy number (which is the problem).
My first source is “North America Data Center Trends H2 2025” by the real estate tracker CBRE.
This quote provides the two little numbers, which are MW of data centers under construction:
“The total amount of new capacity under construction in primary markets declined for the first time since 2020. There were 5,994.4 MW under construction at the end of 2025, down from 6,350.1 MW in 2024. Many planned projects remain delayed due to ongoing permitting, zoning, and power procurement hurdles, underscoring the complexities of scaling infrastructure.”
So, there is only about 6,000 MW of new data centers in the construction pipeline for the entire U.S. That is a tiny number when it comes to the national grid. PJM alone peaks around 150,000 MW, Texas over 80,000, and so on. Combined American peak is over 750,000 MW.
Moreover, installed data center capacity is just around 17,000 MW. While adding 6,000 MW will be a big percentage increase, it will still be an extremely small amount. Note that this is not 6,000 MW a year, because it takes several years to build a data center.
For the huge fantasy number, we go to “Data Centers in the United States” by Clearview.
They have a nifty map showing all the existing and proposed data center clusters.
Here is the big number quote:
“As of April 2026, the U.S. has 602 operating data centers with 16,914 MW of capacity and 889 planned projects that would add 278,842 MW of additional capacity, according to Cleanview’s project tracker.”
Adding 280,000 MW of new power demand would certainly fry the grid, but this is just a fantasy number. These are not “planned projects,” just proposed or merely mentioned by some bigwig.
Unfortunately, it is variations of this fantasy number that are driving the data center panic. Many press articles read as though we are already being swamped with data centers, when the reality is new construction is barely noticeable. CBRE has a nice listing of the latest year-on-year additions for each region. Almost all are tiny.
Note, too, that the tiny amount of demand growth in existing data centers cannot possibly account for the regional surges in electric power prices. This is why there is no correlation between data center growth and power price increases.
It is helpful to see the long-standing parallel fantasy on the power generation side. The queue of applications for wind and solar projects to connect to the grid is over two million MW, when our total generating capacity is just over one million. Almost none of this proposed generating capacity is going to be built, and neither is the fantasy list of proposed data centers.
The data center threat is way overblown, if it exists at all. This supposed threat is just an unfortunate offshoot from the massive AI hype. Untold billions of investor dollars are flowing into the hot AI bubble, so there is tremendous incentive to develop AI data center proposals.
It is possible that over the next 10 to 20 years there will be significant growth in data center capacity. Some states, regions, and localities are already vying for this development. But it will be constrained by available power — in fact, it already is. New dispatchable generation would be good for restoring reliability, and if the data centers want to help pay for that, fine by me.
On the other hand, if the AI bubble bursts there might be relatively little growth. Investment capital can disappear very quickly.
But in no case is there a data center threat to the grid. Data centers are too small and growing too slowly to be a threat.
Thank you for putting some sanity to the myth. The only real threat is to local energy supplies. If you build a data center where there’s little energy feed there’s a problem. But it’s local, recognizable, and can be planned for. Any new construction … industry, homes, businesses …. has the same problem.
The same goes for cooling water, another overblown ‘Threat to the environment/planet/mankind’ (take your pick).!!
AWS was just sued by local community for excessive nitrates in ground water showing up in their wells, nitrates cause health hazards. Nitrates are not used anywhere to treat water to my knowledge (worked in multiple industries with water treatment). Where do the nitrates come from? Farming. Who has deep pockets? Not farmers.
The source of nitrates in the groundwater is probably the fertilization of farmlands.
Several sources of nitrates are home owners lawns, golf course, city parks, and grass athletic fields.
Limited effects. Biggest is more intensive cattle farming, which takes decades to show up
To show up in ground water (wells).
Runoff to surface water can happen fairly quickly.
There are limits to the nitrates allowed in finished (drinking) water.
If the lower limit is exceeded then warnings are issued. Nitrates can interfere with the body use of oxygen. The warnings are aimed at those whose bodies already have a problem with getting enough oxygen. (Babies who are fed formula where it’s mixed with tap water, some elderly, etc.) A healthy person has no problem with nitrates at that level.
Some treatment plants that have a problem, seasonal (from farming) or otherwise, in their source water use zeolite water softeners and even reverse osmosis to handle the nitrates.
As with new generation there is an interconnection approval queue for new demand. This is the biggest constraint on data center development. And if a grid upgrade is required the applicant must pay for it. The panic seems unaware of all this.
Yes. Multiple times I have had to build generation and run as an island until the utility could catch up. Have also had to build substations to utility specifications at my own cost then hand over to the utility to operate (this to avoid schedule que time for the utility to build it).
Suitable utility interconnects or the ability to get them built is a key variable in site selection.
With a little bit of clever negotiation, data centers can be turned into a huge positive for the grid. Data center owners would be happy to build a little extra generation or furnish some critical right of way or pay to loop some interconnects if it helps their business case with less resistance, faster permitting, etc.
“this to avoid schedule que time for the utility to build it”
Que?
Short for Queue
It’s a moot point because Trump is permitting industry to generate their own power. Excess will be sold to the grid increasing capacity not reducing it.
They already had ‘permission’
Regulatory hurdles made it unfeasible. Those have been dramatically reduced.
What changes were made in the Law?
“What changes were made in the Law?”
The endangerment finding has been rescinded but the main changes have been to regulations which are not laws but impediments to productivity.
You forgot what you originally said:”It’s a moot point because Trump is permitting industry to generate their own power. Excess will be sold to the grid increasing capacity not reducing it.”. I asked what Law was passed to accomplish this.
No new law was necessary, just changes in the administration of existing laws. What changes in the law led to the complete closing of the border, something the last President said was impossible? We didn’t need new laws; we needed a new President. And we got one.
Cite the specific regulation.
“I asked what Law was passed to accomplish this.”
No law was needed. It’s being done by the executive branch reducing regulations that Congress never authorized as laws.
No specific law or regulation cited.
“No specific law or regulation cited.”
Reading comprehension is a problem for you. Please re-read replies for understanding. No laws were needed. Regulations were streamlined with permits being issued in as little as a week according to Trump.
The President has no authority to rescind a regulation. That story is another example of a wild Trump imagination — like the reopening of Hormuz.
That is not true. Almost all regulations are issued from the executive branch, i.e., the President. They carry the weight of law, but they are not laws passed by Congress. Only if Congress deliberately and specifically includes a “regulation” in a bill, does its cancelation require approval of Congress.
A regulation based upon evidence can not be overturned wily nilly. There must be evidence that shows cancelation or changes are justified. This can be done solely in the Executive branch at the direction of the President.
It’s not that simple:
Federal regulations are legally rescinded primarily through a new agency rulemaking process under the Administrative Procedure Act (APA), requiring public notice and comment, or via Congress using the Congressional Review Act (CRA). Agencies must provide a reasoned explanation for rescission to survive judicial review, as they cannot simply reverse policy without justification.
Key methods for rescinding federal regulations include:
Agency Notice-and-Comment Rulemaking: Under the APA, agencies must follow the same formal process to revoke a rule as they did to create it. This involves publishing a Notice of Proposed Rulemaking (NPRM), allowing public comment, and publishing a final rescission rule.Congressional Review Act (CRA): Congress can pass a joint resolution of disapproval to nullify a recently issued regulation within 60 legislative days of its submission. Once nullified, the agency cannot issue the same rule again without new statutory authorization.Judicial Review: Federal courts can invalidate regulations if they are found to be unlawful, unconstitutional, or arbitrary and capricious.Executive Orders/Action: A new administration may issue executive orders directing agencies to review, pause, or begin the process of repealing regulations.“Good Cause” Exemption: Agencies can bypass the lengthy notice-and-comment process if they can prove that adhering to it is “impracticable, unnecessary, or contrary to the public interest,” often used for technical corrections or if a court has already ruled the regulation unlawful.Limitations on Rescission:
Arbitrary and Capricious Standard: Courts can reverse rescissions if the agency fails to provide a “reasoned explanation” for the change in policy.Rulemaking Process: Rescinding a regulation is often a time-consuming process that can take months or years.Statutory Authority: The agency must maintain its statutory authority, and Congress can pass legislation that specifically prohibits an agency from using funds to implement a rule.
“It’s not that simple:”
The rescission of Chevron made it that simple.
No, the Supreme Court overruled Chevron deference in Loper Bright Enterprises v. Raimondo (June 2024), meaning the doctrine can no longer be used to justify agency actions. The President cannot rescind any regulation at will; rescissions must still follow the Administrative Procedure Act (APA), requiring reasoned analysis, and now must align with the “best reading” of the law rather than just a “reasonable” agency interpretation
“The President has no authority to rescind a regulation.”
Are you aware that the Chevron defense has been rescinded? Any regulation not specifically written by congress can be rescinded by the executive branch.
No, the Supreme Court overruled Chevron deference in Loper Bright Enterprises v. Raimondo (June 2024), meaning the doctrine can no longer be used to justify agency actions. The President cannot rescind any regulation at will; rescissions must still follow the Administrative Procedure Act (APA), requiring reasoned analysis, and now must align with the “best reading” of the law rather than just a “reasonable” agency interpretation
“The President cannot rescind any regulation at will;”
You don’t understand the implications of Chevron. Regulations that were not specifically authorized by congress are invalid so are rescinded on that basis easily.
Which regulations were ‘streamlined’?
So what coal powered generation has started?. I see the data centres operators looking for nuclear for its caseload.
“So what coal powered generation has started?.”
Where did I say anything about coal?? It can’t compete with natural gas on price so I don’t know why anyone would go that route.
In our region of Northern VA, hundreds of new data centers are connecting to the grid – they just want electrons.
I have yet to see a feasible design for reliably generating 100% of one’s own power.
David,
Have a look at BHP at Newman in Western Australia.
They run their own power station and it supplies their local and more remote mines in the area.
It does happen but this is a typical, (for mining), ‘special’ case because there is no national/state grid near by.
How does BHP get a free pass on CO2 emissions from their power plant and from all the heavy mining machinery? All those CAT heavy haulers with V16 diesel engines used by the miners must emit enormous amounts of CO2.
BTW The humans of Oz exhale 27 million kg of CO2 every day.
And an equal amount is absorbed by plants by photosynthesis.
The issue is that an additional 40 billion tonnes of CO2 is emitted annually by fossil fuel burning in cars and power plants. About half of that 40 billion tonnes is absorbed by the atmosphere, and is responsible for the 50% increase atmospheric CO2 during the industrial age.
Correct. Western Australia’s total generating capacity is around 7.5GW, of which only 5GW maximum peak is required for their main southern SWIS grid. The other 2.5GW is mostly supplied by private power stations in the state’s northern iron ore mines and gas processing plants. This power is mostly derived from local cheap gas.
I have worked on multiple mines worldwide where mines have no choice but to generate their own power using diesel-fueled gensets – there must be at least 1,000 mines worldwide that operate in just that manner.
So it’s definitely not unusual . . . from the steaming wilds of Africa to bleakest frozen Siberia.
How do they get a capacity factor of 100%? Or do they have a bunch of gensets that only run occasionally hence very low CF?
“I have yet to see a feasible design for reliably generating 100% of one’s own power.”
Natural gas power plants can easily be built at factory level scales. Having control of an important input and being able to sell excess power to the grid makes it economically desirable. Trump is claiming permits are being granted in as little as a week.
I assume you mean combined cycle which cannot get close to a CF of 100% and has unplanned outages. You would need standby close full capacity at all times which is very expensive.
Data centers have 100% standby diesel generators even when connected to the grid.
They will likely build them over capacity for future expansion potential because of the upfront costs and the fact they can sell excess power to the grid. Weather outages are not that common. Backup generators are less costly than lost production in most cases.
It’s done all the time in the oil and gas industry. I suspect mining too.
How? What is the design that assures no outages? I suspect oil, gas and mining just accept some outages, planned and unplanned.
The sugarcane industry has been doing it since 1911. Sugarcane bagasse is the fibrous byproduct of crushing stalks, which sugar mills burn in boilers to generate heat and electricity, making many plants entirely self-powered or even net exporters of energy
With no outages, planned and unplanned? The turbines run 100% with no downtime? I do not think so.
Do the Feds have this authority? Interconnection is up to the utility and ISO where there is one. FERC does not approve individual hookups.
EPA hurdles have been eliminated. What utility would not welcome reliable extra power?
Which ‘hurdles’ have been changed?
“Which ‘hurdles’ have been changed?”
“The EPA has eliminated several regulatory hurdles, including the reconsideration of regulations on power plants and the repeal of the Clean Power Plan, which previously set emission guidelines for states to limit carbon dioxide emissions from existing power plants.”
There are no hurdles. Utilities prefer sosl and wind–about 90% of new generation is solar and wind because they are cheaper and faster to site and build.
“about 90% of new generation is solar and wind because they are cheaper and faster to site and build.”
You don’t live in the real world.
Data From FERC and the EIA:
“Renewables accounted for roughly 88% of all new U.S. electrical generating capacity added in the first 11 months of 2025, driven heavily by solar, which provided over 72% of new capacity
. Globally, renewable energy grew to nearly 50% of total electricity capacity by early 2026, with solar leading the expansion
U.S. New Generating Capacity Trends (2025-2026):
Global Renewable Expansion:
Key Drivers:
Note: Data from Federal Energy Regulatory Commission (FERC) and U.S. Energy Information Administration (EIA) usually focuses on utility-scale (>1 MW), but small-scale solar (rooftop) adds significant additional capacity
.
If a business wants to connect to an interstate pipeline it would be an extension of that pipeline. FERC approval is required for the construction or extension of any facilities used for interstate transportation.
State requirements are all over the board. Six states require no approval at all to connect to a pipeline, build your own power plant, and produce your own energy.
If the pipelines are delivery gas, you still need to have a contract for gas supply to run the generators. You don’t want to depend on spot market prices and volumes.
No can do if the gas supply is already sold. Its like water, a pipeline can only deliver what comes in at the beginning
The first discrepancy I noticed is that the planned 889 is close to the existing 602, only 50% higher, while the planned 278,842 is far higher than the existing 16,914 — closer to 20 times as high at first guess, but actually only 16.5 times as high. Far more than the 1.5 times by count.
The existing data centers come to 28 MW per data center. The planned data centers come to 313 MW per data center, 11 times as much per data center. That doesn’t pass my smell test. It implies that AI data centers are 11 times as big as old fuddy duddy data centers.
From what I’ve read about the increasing power per rack for data centers, the 11X power requirement for new data centers is not much of an exaggeration.
Could be that AI chips are energy hogs. Another interesting thing is that conventional centers tend to be steady consumers but AI demand can be erratic. There is a lot that needs to be better understood and I found it hard to find.
For a given operation, e.g. multiplying two numbers, the A.I. chips are relatively efficient. The energy consumption comes from A.I. machines doing vast numbers of operations in parallel. This allow for a lot more processing in a given volume, which translates into more power consumption for a given volume.
The big consumer of processing power has been building models, which does not need to be done in real time. This could allow the A.I. data centers to reduce power consumption in times of peak system demand, i.e. acting as an interruptible customer, which would reduce the requirements for new generation capacity.
Between 700-1100 W per processor. Way ahead of what ordinary cpus use
They are . Little monsters compared the previous processors which weren’t to different from a commercial grade CPU. The price for a set of say 6 is also eye watering
On top of that the AI data centers overall are much much larger in area. Some of that is planned for say a decade ahead so not needed right now
Nvidia is well along in construction of the Aurora AI Factory (Nvidia calls data centers “factories”) near where I live in Manassas, Virginia. It is gigantic, and will consume 98 MW. It will, among other things, serve as a test bed for what they are planning as the next step: gigawatt data centers.
Don’t be surprised if the gigawatt data center is built in West Virginia. The political climate in Virginia for a big energy consumer is a huge problem unless you can wait a decade or two and be happy with whatever reliability offshore windfarms can provide.
West Virginia is offering data centers a lot of perks to locate there, especially if they use power provided by coal burning power plants.
Land tends to be a lot cheaper in W VA as well.
It does make sense. It’s all about scale, and to a lesser degree economies of scale. 279MW sounds like a lot, but it really isn’t. For example, a Siemens SGT5-9000HL can do almost 600 MWe simple cycle and almost 900 in combined cycle. Even a lowly Frame 7 FA can do 200 MWe.
Turbines would probably be installed in triplets for reliability. One running, one Hot/rotating standby and one cold standby/maintenance unit.
Yes triplets is the design I keep seeing but surely this is very expensive compared to utility industrial power rates which are very low.
AI requires a huge amount of power. They retrieve, store, and analyze huge amounts of data at high data rates. That’s why Microsoft is working with Constellation Energy to restart the Three Mile Island nuclear facilities.
Because they consume huge amounts of power they generate a huge amount of waste heat. They use various forms of water cooling, which creates potential problems with their water consumption.
It’s also the reason why a small company with six-figure revenues, but a product in the commercialization stage of an unproven technology, has a market cap of $1.87 Billion dollars. Their technology would reduce power requirements for AI data centers by an estimated 30%. Power is a big budget item.
Test regular text …. bold on … bold off … Italics on … Italics off
It’s everywhere — stories, comments. Well, not every everywhere, “Post Comment” isn’t bold and a few others.
Oh, so it’s not just me, huh? 🙂
Yes, somebody changed the stylesheet.
I think it’s easier to read, given my failing eyes. 👓
re: “I think it’s easier to read, given my failing eyes.”
I went through a phase where I required background screen and text to be inverted (screen background black, text white) in order to read anything off my PC … I experienced fast-onset cataracts (owing to circumstances I won’t describe) but since corrected with surgery (lens replacement) and have reverted to normal screen scheme. If you are having difficulty with something similar this trick on a PC or Mac or tablet I can recommend that screen/text inversion might work for you as well.
Along with your good suggestion, I use browser-specific reading options. For example, Firefox has a reader view that provides a clean, readable layout with adjustable fonts and backgrounds. The Brave browser’s speedreader option is similar.
Coincidentally, I’m scheduled for an evaluation for cataract correction in a few weeks. In my case, the cataract appears to be congenital.
What is the name of letter font? The letters in this comment box are lighter, but become darker after the comment is posted.
The font family is Segoe UI Semibold. The reason it looks bolder is because the font weight is set to 600.
Slight correction: On Windows browsers it’s likely the Segoe. On MacOS it’s probably Helvetica or San Francisco.
This simplistic narrative ignores the substantial problem of the concentration in data center growth. For example in my northern Virginia county the world’s largest data center campus is being considered by the courts for approval. In addition , data centers are proliferating in other areas of our county. The local utility (the largest electric cooperative in the US) projects their load will continue to increase and by 2035, increase their total demand 10 fold. This is one hell of a challenge. In addition, water demand for data center cooling in the Potomac basin is projected to risk severe water shortages by 2030. So, no you cannot use macro level data for the US to dismiss a real problem for local communities in our region .
“10 fold” in 9 years? Sets off my BS detector. Someone’s scare-mongering.
The utility is NOVEC. give them a call and ask them. Until you do, you have no argument
On the contrary the supply chain needed is likely impossible. CBRE says that growth through 2030 is already set and it is not huge. Most of these proposed projects do not even have funding or real estate.
Not correct. Our local area already has commitments for 33 million sq feet of data centers (a good share of the 33 million’s already built) and another 33million are on its way
And look how many commitments there have been for wind farms and solar farms, which never came to fruition. I’ll believe it when they are actually running, not just from what someone saw in a press release.
Sorry, I live here. And they are up and runnng everywhere.
AI generated response:
AI data centers are being constructed at a historic, record-breaking pace in the US, with construction spending in early 2026 reaching an all-time high of $25.2 billion in January alone. Over 1,500 new data centers are currently in various stages of development, with over 15.9 GW of IT capacity under construction as of March 2026. Texas and Virginia lead the nation with 140 and 136 data centers under construction, respectively.”
But:
“As many as 40% of US data center projects planned for 2026 are expected to miss their completion dates by more than three months, according to satellite imagery analysis.
Despite the rapid construction, the sector is facing severe supply chain constraints that are causing significant project delays.”
My cited source which track projects says just under 6 GW under construction. What is your AI bot’s source for 15.9? Which is still nothing compared to the 260 GW supposedly planned.
I have the BS smell test argument. It’s your choice to pretend that 10-fold in 9 years is plausible. It’s my choice to believe that you copied the wrong information, misinterpreted it, or something else. Do you also clap for Tinkerbell?
This was announced by the President of the Utility to a live audience.
Instead of insulting someone, why not simply do a fast search, yourself.
You might need a new nose.
“The Northern Virginia Electric Cooperative (NOVEC) is predicting a massive increase in power demand, driven almost entirely by the rapid expansion of AI data centers in its service area.
Total Power Capacity: NOVEC expects its data center load to reach 12 gigawatts (GW) by 2035. This represents a twelvefold increase from its current levels.Customer and Building Growth: To reach this capacity, NOVEC anticipates powering approximately 178 data center buildings by 2035, up from the 58 buildings it currently serves.”Moreover
“Dominion Energy’s forecasts for Northern Virginia also predict an unprecedented “supercycle” of growth, with expectations that overall power demand in its delivery zone will double by 2039. While NOVEC’s predicted 12 GW of growth by 2035 is massive for a cooperative, Dominion remains the primary utility for the region and is planning for the largest surge in demand since World War II.”
Dominion will “only” double, but that’s because it serves a huge area, now, with an already large base. The data center growth is predicted to be 9 GW, very similar to NOVEC’s forecast of 12.
The first link includes local growth statistics. Northern VA was biggest but still only grew 1,000 MW last year. These mammoth projections are assuming a lot of these projects actually get built which may be unlikely. But as I said if the utilities want to build a lot of new generation and restore reliability that is fine by me.
Me too.
The problem is the Green New Deal and renewables have made the power grid unreliable and unable to carry the needed power. We need reliable, affordable, dispatchable power and it requires a backbone of nuclear plants for base power plus gas and coal plants that can be spun up quickly on demand.
There was no GND policy ever.
It was called the Inflation Reduction Act.
No, that was not the GND.
It was the GND wrapped in a different label because by the time the bill was passed, the label “Green New Deal” had become toxic. Therefore, it needed to be called something else in order for you wackadoodles to appear somewhat coherent.
Nope. Wasn’t connected in any way to the GND. I know. Because I lobbied for the IRA.
From someone who thinks Lindzen and Happer are irrelevant, I believe that you believe that.
Linden and Happer have never published a peer reviewed scientific paper disputing climate change or its cause. Presumably because they are no longer competent.
Please explain to us all why you think LindZen and Happer are not competent.
They’ve never been able to publish a peer reviewed scientific paper disputing the body of science on climate change and its cause. Yet they claim the body of science is wrong.
global_warming_what_is_it_all_about.pdf https://share.google/sDk8KTJf4RKcFglEQ
Not published on a peer reviewed scientific journal. So not part of the body of science. Ie, it’s fraud
Thanks for showing that you are logic challenged – that is a non-sequetor – you conclusion DOES NOT follow from you data.
‘Peer review’ is synonymous with censorship. It is not in any way related to the scientific method, and as Happer himself has said, the vast majority of ‘peer review’ does not name the ‘reviewers’ or how and by whom the research was funded. Which explains why the ‘scientific’ journals publish so much claptrap.
Censorship refers to government prohibiting speech. Peer review simply means review by your scientist peers to screen for errors in math, logic, or supporting data. Happen and Lindzen can’t seem to pass this simple screen.
You do realize that many peer-reviewed articles have been withdrawn for being incorrect, right? That tells me that peer review is not a proper bar to use in judging scientific relevance.
That tells me that Deniers don’t understand the scientific disciplines
Was Einstein’s papers fraudulent since they weren’t “peer reviewed”? How about Max Planck’s seminal works, they weren’t peer reviewed?
How about all the gas laws from Boyle’s, Charles`, Gay-Lussac?
I could go on. Are these all fraudulent? If not, then peer-review is not the correct baseline to prove/disprove scientific relevance. You need to find another horse to flog.
Peer review was not common practice in early 20th-century physics. Journals were less crowded, and editors often focused on rejecting only “crackpot” work rather than rigorously vetting new ideas.
today, peer review still separates incompetent Deniers from real scientists
This is but one paper published by Lindzen disputing the erroneous conclusions of climate alarmism. If you want to find more, look them up yourself.
It’s not a paper. It’s a slide show.
Whatever, Warren, Keep chugging the Kool Aid.
You don’t submit a slide show to nature or Science Journals for publication as a ‘paper.
Hey genius, Warren, please debunk this:
EACH OF THESE FACTS DEBUNKS AL GORE’S CLIMATE SCAM:
1-There is NOTHING UNUSUAL about our climate – The Holocene (Our current inter-glacial) has been both warmer and cooler than now BEFORE man emitted CO2 See:
http://www.sustainableoregon.com/ipcc_says.html
An Estimate of The Centennial Variability of Global Temperatures, Philip J. Lloyd, DOI: 10.1260/0958-305X.26.3.417, http://multi-science.atypon.com/doi/abs/10.1260/0958-305X.26.3.417
http://www.debunkingclimate.com/natural_climate.html
2-Global warming started 200 years BEFORE man started releasing CO2
See: https://www.ssb.no/en/natur-og-miljo/forurensning-og-klima/artikler/to-what-extent-are-temperature-levels-changing-due-to-greenhouse-gas-emissions/_/attachment/inline/5a3f4a9b-3bc3-4988-9579-9fea82944264:f63064594b9225f9d7dc458b0b70a646baec3339/DP1007.pdf
3- CO2 changes FOLLOW, NOT LEAD, temperature changes in the ice core data AND at all other times. See:
https://judithcurry.com/2023/09/26/causality-and-climate/
4- NO ONE HAS EVER shown good evidence that man’s CO2 is causing serious global warming. (prove this wrong by posting actual evidence that man’s CO2 is causing serious global warming.)
5- Solar cycles are a better fit to climate than CO2, thus negating the claim that the simultaneous rise of temperature and CO2 proves CO2 is causing increased temperature. See: http://www.sustainableoregon.com/CO2_Solar_Corrlations.html
6- Recent warming is at the same rate as the late 1800s but now with much more of man’s CO2. (More of a cause should cause more effect.)
http://www.debunkingclimate.com/co2_rate_of_warming.html http://www.debunkingclimate.com/no-rapid-warming.html#no-rapid-warming
7- Most climate records start at the end of the coldest time in 8,000 years, so natural warming is the best explanation for our current warming: “The Little Ice Age (LIA), which lasted from about 1250 to 1860 AD, was likely the coldest period of the last 8000 years.” from: https://www.sciencedirect.com/science/article/abs/pii/S0277379122001627
8-We have never had accurate whole earth coverage of temperature until satellites in 1979, so it is not possible to make any claims about today’s climate being unusual compared to meaningful history. http://www.debunkingclimate.com/lack_of_data.html
9- Human CO2 release warms the climate less than 0.03?C
https://arxiv.org/pdf/2304.01245.pdf
10- Melting glaciers are uncovering artifacts from Roman and Medieval times, proving that those areas were ice free during those times and thus were warmer than now, which in turn, proves that our climate is not unprecedented which proves that man’s CO2 is not having a major (if any) effect on climate. See:
http://www.debunkingclimate.com/glaciers.html
https://notrickszone.com/2019/08/26/a-1000-year-old-forest-buried-under-alaskas-mendenhall-glacier-uncovers-a-warm-medieval-period/
https://climateaudit.org/2005/11/18/archaeological-finds-in-retreating-swiss-glacier/
See DebunkingClimate.com for real evidence.
Feel free to disagree by showing actual evidence that man’s CO2 is causing serious global warming. (Or show your unwillingness to learn by posting a laughter emoji.)
Please note:
1-Evidence of warming, unusual weather, storms, floods IS NOT evidence that man’s CO2 is the cause.
2-Correlation is not causation http://www.debunkingclimate.com/CO2_Solar_Corrlations.html
3-An expert’s assertion, government’s assertion, consensus of experts, polls or majority belief are not evidence. They are hearsay.
4-Climate models are not evidence.
5-Warmest weather in 100 years means it was warmer 100 years ago when CO2 was lower.
6-If an event is NOT unprecedented, then you have to explain why whatever caused the earlier events is NOT the cause of the latest occurrence of that event.
Evidence is actual data PRO AND CON with reasoned analysis and logical conclusions while FULLY CONSIDERING OPPOSING evidence.
When do we expect to see your “debunking” published in a reliable peer reviewed scientific journal? (My guess is never)
David, since YOU are so SMART,
Please show us ACTUAL EVIDENCE that man’s CO2 is causing serious (not trivial) global warming.
PLEASE – NO EVIDENCE of warming, we already know the Earth appears to be warming after the coldest time in 8000 years. AND NO LESSONS on how the climate works – we know that too.
I AM ONLY ASKING FOR REAL EVIDENCE THAT IT IS CAUSED BY MAN’S CO2.
Note to readers:
I have asked this question (asking for evidence) many times and have NEVER received a response that is not debunked by a simple question or two such as:
1–“how does the fact of warming (or bad weather, etc.) prove that man’s CO2 is the cause? OR
2– Please explain what caused Minoan, Roman and Medieval warm periods to be warmer than now without man’s CO2 AND to occur every 1000 years, with our current warm period fitting that regular cycle
That tells us that most climate zealots DO NOT ACTUALLY know of any evidence and didn’t realize it until asked to show it. Plus a few others who are lying and they know it.
-30-
That question should be addressed to Warren.
I suggest further reading and perhaps a university level course in atmospheric science. MIT has several online.
“The history of CO2 and climate shows that carbon dioxide is a primary driver of Earth’s temperature, with current human-induced levels (over 420 ppm) rising 100–250 times faster than natural post-ice age increases. NASA Science (.gov) +1
. Since the 19th-century discovery of the greenhouse effect, scientific understanding has evolved from identifying CO2 as a climate factor to proving that industrial emissions are causing rapid, unprecedented global warming. AIP.ORG +1
.Early Scientific Discoveries (19th – Early 20th Century)
and water vapor as the gases responsible for this heat absorption. Svante Arrhenius later calculated that cutting CO2 levels in half could cool the planet, triggering an ice age, and suggested that burning coal could cause future warming. AIP.ORG +1
Discovery of Rapid Rise (1950s–1970s)
Modern Understanding (1980s–Present)
Historical Climate Anomalies
Key Takeaways
“The history of CO2 and climate shows that carbon dioxide is a primary driver of Earth’s temperature, with current human-induced levels (over 420 ppm) rising 100–250 times faster than natural post-ice age increases. NASA Science (.gov) +1
Is that from a scientific peer reviewed paper? If it isn’t, then it doesn’t meet your own self-imposed bar for proof. Try again!
its a verifiable fact. Do you have a peer reviewed source that disagrees?
[“The history of CO2 and climate shows that carbon dioxide is a primary driver of Earth’s temperature,“]
No, it isn’t …
Simplistically, the Sun is the primary driver !!
And water vapour is the thermostat; CO2 only plays a minor role.
You need to take a course on atmospheric thermodynamics.
Since the suns output slightly diminished over the last 50 years, while earths temperature rose rapidly, that would seem to discredit your argument
Since the suns output slightly diminished over the last 50 years, while earths temperature rose rapidly, that would seem to discredit your argument.
What about the study referenced here a couple of days ago?
One of the fallacies of a warming atmosphere is the problem with lumping two different temperature functions, Tmax and Tmin, into one simplistic average.
Tmin has been the primary driver for land temperatures. This has reduced the diurnal range which has wreaked havoc with the time series trending.
Another is ignoring the fact that the driver for the atmosphere is the surface, land and ocean, and not an immediate reaction to the sun’s radiation.
Are u denying that the planet is warming?
Are you denying that it is the minimum temperatures that are rising and not the maximum temperatures?
The data shows both are rising more rapidly than at any time in millennia. Once again the question: since the suns output fell slightly in the last 50 years, how can the sun be the cause of this rapid warming?
Oh, no you’ve told a falsifiable porkie. You just shot your credibility down the drain.
Google’s first link was Wikipedia.
Exactly, No policy was actually passed in the legislature. So wasn’t implemented.
“Note that this is not 6,000 MW a year, because it takes several years to build a data center.”
Absolutely wrong. I started with a DC less then 1.5 years ago. At only one campus when I began the first building was just starting to accept racks and the second buildings shell was up. Today the 4th building is now accepting racks and the 5th building shell is up (final build on that campus). Just around the corner they are prepping ground at yet another campus right now. This is going on all over the geographic area I live in. An entire campus worth of buildings takes 2 years not 2 years per building.
I think I read it typically takes 1.5 to 2 years for construction and commissioning if there are no delays which there often are.
Trust me it’s going much faster then that at least in my area. Before seeing it for myself I would of thought 2 years would be a reasonable target.
I think the article is misleading. The issue is local energy use and allocation, not overall grid power use.
Simply because on a local level real choices need to be made between supply and demand as well as the grid structure.
And this is what is going on right NOW.
Combine that with increased electrification of just about everything and it is pretty clear..
That it is a matter of local peak demand not grid power use is my point. To date that has been very small.
The data center ‘projections’ are a result of the AI hype, which strongly reminds me of the ‘dot com’ bubble that burst around 2000 and took down hyped companies like Sun Microsystems, whose then advertising was “we put the dot in .com”. The Sun wreckage was sold to Oracle for a pittance.
And to the extent AI does experience strong growth, there are at least four technical innovations on the horizon that will significantly lessen electricity demand.
AI is not much of a factor. Growth in our area is due to crypto, streaming, transactions, and conventional data demand
Well said, thanks!
Thanks Rud
Some of the older centres will be replaced. Then there is the separation of storage and search, and the novelty factor.
Is it true that the latest Chinese chips use less power per task ?
While a single instance of consumer functions uses vastly less power than a training run, their cumulative lifetime energy consumption is significantly higher, often accounting for 80% to 90% of a model’s total energy footprint.
The AI bubble will burst because the LLMs won’t lead us to Artificial General Intelligence. The LLMs are getting better but it’s the wrong direction.
AI isn’t going away, however. The LLMs have useful applications.
BTW, I hate that we use the term “AI” to refer to these LLMs. They aren’t intelligent.
I wish the author had addressed the water issue and why it’s not really an issue.
That’s a great topic for a post — how do data centers use water, how the water is recycled, etc.
The water is evaporated, not recycled
Can be either, or can be purely mechanical cooling (refrigeration), if there is enough power available.
Doesn’t matter. Any heat rejected by water cooling., refrigerated or not, is transferred to the ambient air by evaporation.
WRONG!!
Look at your car for how its done!
Water cooled cars DO NOT COOL by evaporation.
Terminology. cars rely on air cooling as the last heat exchange step from the engine to the atmosphere. Water is used to transfer heat from the engine to the air cooled radiator. The term water cooling as applied to data centers refers to the use of water cooled chillers which in turn use cooling towers to evaporate water in order to exchange heat between the water cooled chillers and the ambient air.
You are talking about old school data centres … new direct cooling data centres work exactly the same as a car radiator.
There advantages
1.) Higher efficiency
2.) Water is in closed loop
3.) No dust intake
https://www.computacenter.com/who-we-are/blogs/direct-liquid-cooling–the-new-gold-standard-for-data-centers
The whole water argument will soon disappear it’s like arguing cars will use up all the water because it’s in their radiators.
Your link is talking about direct to chip cooling which efficiently transfers heat from the chips to the data center’s refrigeration system. It is certainly a superior method of intermediate heat exchange vs older methods of air cooling the chips.
However, I am referring to something else –the final (not the intermediate) step in the multistep process of moving heat from chips to the ambient air outside the data center. Once the heat is moved from the chips to a liquid, then to a chiller, the chiller’s refrigerant condenser must then be cooled either by air (relatively high energy consumption, but no water used), or by water which is subsequently evaporated in a cooling tower (energy efficient, but high water consumption).
Most modern large scale data centers in moderate or warm climates use the latter method of final stage water cooling and evaporation. Direct to chip liquid cooling can be used regardless of whether the data center is ultimately refrigerated by air cooled chillers or water cooled chillers.
OMG, seems Beeton has never heard of condensation tubes.
Has never seen air passed through a car radiator of cooling.
Servers can be cooled by the equivalent of air-conditioning, with no water needed at all.
I have a water-cooled CPU in one of my computers.. No evaporation required.
Server room at a place I used to work, had 20 sets of racks, all cooled by standard airconditioning.
Your knowledge on the issue seems to be basically zero. !!!
Of course a cars radiator is air cooled, and so ‘air cooling’ is the final heat exchange step to cool a cars engine. Water is only a medium to move heat from the engine to the radiator.
For data centers, the term ‘water cooled” refers to the evaporation of water in cooling towers which cools water circulating in the refrigerant condensers of a chiller. For large data centers in moderate or warm climates, this is the predominate cooling method. Alternatively, a data center chiller can use air cooled condensers, which reject heat to ambient air in a sensible cooling process, ie, no water is consumed .
Above, you said “Any heat rejected by water cooling., refrigerated or not, is transferred to the ambient air by evaporation.”
Which you now admit was totally wrong.. Thanks 🙂
Stop digging.. you are in way over your head already , and your BS is collapsing in on you.
“Any heat rejected by water cooling, refrigerated or not, is transferred to the ambient air by evaporation.” from my first post is equivalent to “For data centers, the term ‘water cooled” refers to the evaporation of water in cooling towers which cools water circulating in the refrigerant condensers of a chiller.” from my last post.
What are you arguing about, BNICE2000?
Evaporated water usually falls back to earth somewhere. It will be recycled maybe not locally. Maybe not quickly.
Water is used for evaporative cooling. In humid areas you can use chillers systems that are then cooled using cooling towers. In dry areas they can use direct evaporative cooling (essentially giant swamp coolers) to cool spaces. In either case they go through a lot of water. Not all the water will be used as it concentrates whatever is in the water (minerals mainly) so the sumps will be regularly flushed to prevent mineral buildup. The concentrated water can be either dumped or returned to the utilities depending on local setup and regulations. Manufacturing and DC’s in the area I work send water back to the utility.
WRONG!! —
Look at your car for how its done!
Water cooled cars DO NOT COOL by evaporation.
WRONG!!–
You can use water or refrigerant to air coolers but cooling a large data center off refrigerant to air coolers would take up a huge foot print and I’m actually not sure it’s possible. Chiller + Cooling towers will give a much larger cooling density without the huge footprint. As an example one 500 ton chiller would need to be replaced by 20 25 ton units and it will take more than one 500 ton chiller. You are seriously underestimating exactly how much cooling is need for a DC. I sure as hell can’t give specifics without endangering my job but cooling is well north of 500k tons of cooling per building.
Actually, many data centers in our region have air cooled chillers mounted on the roof or alongside the building. Water cooled is dominant among the largest data centers, but air cooled chillers are still common.
The 280,000 MW of load that the industry talks about wanting to build likely is a fantasy but half that will be beyond what the system can manage and the billions of dollars seeking an advantage in the AI race are very real. Looking backward at traditional development rates is unlikely to yield clarity. Musk is likely the fastest but he built the 100 MW facility in Memphis in 8 months. His competitors are not sitting idly by. AI will push every limit in the development of capacity in the next few years.
Have you all seen the community opposition organizations like this one?
https://protectroundrock.org/datacenters.html
The real question is whose money is being used to gin up all this opposition?
Russia and China plus a few rich idiots come to mind.
see: http://www.debunkingclimate.com/russia-articles.html
There’s plenty of opposition from voters in Northern VA.
https://robertbryce.substack.com/p/ai-rejected-tracking-the-great-data
Very nice.
But, but it makes such an easy target for deflecting the real issue of bad public policy in blue states.
Why is it that the states leading in wind power are iowa and texas?