About Those Data Centers. Yeah, the Ones with Targets on Satellite Imagery. Those Data Centers
May 4, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I read “Amazon Stuck with Months of Repairs After Drone Strikes on Data Centers.” The write up contained a number of factoids I found interesting. One example is the statement “AWS stops billing Middle East cloud customers as repairs to war damage drag on.”
Let’s think about this statement, assuming that it indeed accurate.
Thanks, Venice.ai. Good enough for an act of generosity.
Amazon, for some in the conflict zone and peripheral areas, synonymous with cloud computing. I would argue that in certain areas Microsoft Azure has a more prominent profile. For the present discussion, let’s focus on Amazon. The company does not publish a name, location, and facility capability description for each of its infrastructure locations in Bahrain, UAE, Saudi Arabia,Oman, and Qatar. But Amazon lists AWS regions and availability zones and direct connect locations, usually providing the city and the country. With that information, it is not too difficult to pinpoint the exact location of Amazon facilities by [a] asking a taxi driver if one is in country, [b] making some contact with government-linked business “facilitator,” or [c] getting some gig work market research via Fiverr.com.
My point is that the article makes clear that at this point in time (May 2, 2026 at 236 pm US Eastern time) Amazon has a number of facilities that could become targets at any time. This adds to the challenge Amazon and other data center operators face in contentious parts of the world. How can these facilities defend themselves against drone attacks? If attacked, how can the damaged facility be repaired and returned to “normal” operation?
The reason I bring this up is that the cited article makes a big deal about Amazon not billing Middle East cloud customers for the damage that has occurred. That’s okay. There are some problems that plague the approach:
- So far Amazon has been able to cope with the damage to the handful of its data centers; however, the implication in the write up in my opinion is that getting these facilities back to pre-attack status is not quick and easy. Those two factors translate into costs. Amazon is investing in AI; the company is terminating some employees who are not productive. The “grace period” cannot be extended indefinitely it seems to me.
- If hostilities flare up, data centers are easily identified without help from nations with satellite imagery. This means that a coordinated attack could take out a number of data centers simultaneously. I can envision war planners thinking, “Why not take out the Bahrain and Saudi facilities on one day and then go after the Microsoft Azure facilities the next day?”
- Companies providing services to the large data center operators can also be easy and vulnerable targets. What defensive measures does Equinix-type companies have in place?
Let’s think about how commercial firms like Amazon can protect these quite large, easily findable, and highly vulnerable data centers. Options range from the use of World War II methods like sandbags, temporary defense shielding, and Kevlar shrouds. These are cheap and visible “protections.” The problem is that a drone only has to fly into a power or cooling vent, and the “off switch” can be flipped.
More robust defensive measures run into the same problems as the repairs to damaged facilities: Time, resources, and costs. Data centers vulnerable today will be vulnerable to some indeterminate time in the future. Thus, until the facilities are hardened, they are at risk.
If we think in terms of an aggressor attacking data centers owned by US companies in other countries, the attack surface becomes much larger.
Several observations are warranted:
- The success of attacks on the Amazon facilities makes clear that these structures were not designed, engineered, and constructed to withstand a drone- or other kinetic attack. That seems to have been what I would call an ill-considered decision. The reasons may range from keeping costs down or MBAs in the US don’t think too much about asymmetric issues.
- Data centers now under construction may require expensive hardening if these facilities are located in regions where they become large, easily located targets. Stated another way: The cost of those under construction data centers are likely to go up and require more time before they go online and produce revenue for their operators.
- Future data centers will warrant some design, engineering, and construction scrutiny. The success of Ukraine drone in damaging distant Russian infrastructure is a new factor at least for the “Z” folks.
The Ars Technica write up includes this passage:
The latest AWS status update comes just after another data center developer, the London-based Pure Data Centre Group, said it will pause Middle East data center investments until the ongoing Middle East conflict subsides.
Pure Data is thinking clearly. However, can one really go home again or step into the same river twice? The massive US investment in data centers makes these facilities even more attractive targets than they were when Amazon experienced its first attacks. Going forward, addressing the issue means data centers are going to get more expensive, not less expensive. Statements like “stop billing” might be tough to honor.
Stephen E Arnold, May 4, 2026
The EU: Another Government Does Some Fancy Dancing
April 30, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold. I find it interesting that AI detectors identify my writing style as AI output. I suppose I should be flattered, but I just don’t care.
The European Union loves to fine American companies. I suppose if I could ring up some notional money, publicize my stance against American overreach, and get elected — I would do some fancy dancing.
I read an article or exposé published by Investigate Europe. “How Big Tech Wrote Secrecy into EU Law to Hide Data Centers Environmental Toll” states:
Microsoft and DigitalEurope, a lobby group whose members include Amazon, Google and Meta, secured a secrecy provision in EU law to block public access to critical information on data centers’ environmental impact, Investigate Europe can reveal.
What’s a Digital Europe? My recollection is that it consists of more than 100 high-tech outfits and what are called “national trade associations.” I interpret this phrase to mean “soft advocates” for big technology.
Thanks, Venice.ai. Good enough.
Investigate Europe figured out that big AI tech outfits or what I call BAITs worked hard to make sure that the environmental impact of large-scale data centers was not widely known, disseminated, talked about, or provided to citizen groups willing to carry signs. No big surprise to me. Investigate Europe, on the other hand, appears to have be in a cloud of unknowing. I find that darned interesting.
The write up reports:
With the EU set to triple its data center capacity in the next five years, the European Commission started collecting key metrics like energy efficiency and water consumption from facilities. However, information on individual facilities’ footprint is kept secret, after industry pushed to amend the 2024 legislation to classify it as confidential and commercially sensitive.
This strikes me as standard operational practices for BAIT-type outfits. Now folks are worried that bad things will come about when the data centers power up. The write up says:
Europe is building data centers at break-neck speed, with €176 billion in investment expected over the next five years. The rush has triggered widespread concerns about pollution and intense energy use as well as impacts on communities and natural habitats.
Let’s not forget consumer impacts like “power shaping.” Someone’s sensitive electronic devices, including a few with uninterruptable power supplies or back up generators, may be tested. Failures mean that someone might hear, “Daddy, my mobile is not charging.”
Who crafted the confidentiality wordage? I don’t know, but it seems as if some of the Microsofties played a role. The report offers this factoid as a quote from Bram Vranken at the Corporate Europe Observatory, an NGO in Brussels:
“The fact that the Commission copy-pasted a Microsoft amendment is shocking,” Vranken said. “Who does the Commission really represent: Big Tech or the public interest?”
My answer to this question is that dolphins, herring, and black stork are probably not the EU’s primary concern when it comes to data centers.
Investigate Europe is working with other publishing outfits to pass the word that fancy dancing has taken place. More hoe-downs are likely to be held to make sure that power, emissions, noise, and assorted infrastructure projects are kept under wraps.
This is not a surprise.
Stephen E Arnold, April 29, 2026
Give Up And Learn To Love Data Centers
April 20, 2026
Tech Radar shares one story that puts a damper on any good news related to AI technology going carbon neutral or green: “An Ex-Programmer’s Devastating Take On AI Data Centers Is Going Viral — And It’s Hard To Ignore.”
Data centers are also bad for the environment (land and water), takes advantage of tax breaks, and brings few jobs to stimulate local economies. Ravenna, Ohio is a small town that might become home to a new AI data center. An ex-programmer by the name of Hollingsworth delivered an insightful speech about the harms of datacenters at a meeting in Ravenna.
Hollingsworth challenges the idea of closed-loop water systems and believes that the forever cooling chemicals will probably make their way back to the water table despite all the studies that say otherwise. Hollingsworth said:
“Ohio is becoming a major battleground over AI infrastructure, with communities questioning whether the trade-offs are worth it. The push back feels less reactionary from somebody like Hollingsworth because of his credibility as an ex-programmer and working with AI models in the past. He frequently says that he is not against new technology or AI in particular, but the costs required for it to run come at a price, and it’s one that the city and its residents will end up paying if the planning permission is given the go-ahead.”
Some folks are a bit more feisty. “Facing Angry Residents, Minnesota City Halls Are Ground Zero for Data Center Battles” reported:
Just a few years ago, neither cities nor residents had much reason to think about data centers. The facilities have operated at a small scale across the state for decades, often treated like warehouses or light industry. But as AI fuels a data center boom, Minnesota cities face a seemingly exponential number of development proposals without much state regulatory guidance. And residents’ concern over the large facilities and their societal effects are rising just as fast. The charged debate touches on everything from energy and water use to noise and legal risk. City officials are caught in the middle. Some, such as Bloomington and Golden Valley, have passed zoning ordinances to set rules on location, noise and size. Others, including Eagan, Carver and this week, New Brighton, have chosen one-year pauses on new data centers to study their impacts.
My view is that BAIT (big AI tech) outfits will hire legions of attorneys, support political candidates who like the idea of data centers, and promise a green approach. States like Ohio might roll over and let the BAIT outfits scratch their tummies. Other states, possibly Minnesota, might take a different approach.
What’s interesting is that data centers take time to build. Certain components are in short supply and expensive. The permits can be a hassle. But the big issue is technology advances. More efficient algorithms may reduce the need for the data center build out. But the dark thoughts might be: What if AI demand does not meet the projections of the spreadsheet jockeys? What if different methods deliver significant efficiencies? What if someone comes along with a way to leapfrog the Google transformer method.
Poof. The data center problem declines in important. Yep, dark thoughts.
Whitney Grace, April 20, 2026
Oracle Chained: Is James Bond Available?
March 18, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I read a data center report with an ominous vibe. “Oracle Is Building Yesterday’s Data Centers with Tomorrow’s Debt” reports:
OpenAI is no longer planning to expand its partnership with Oracle in Abilene, Texas, home to the Stargate data center, because it wants clusters with newer generations of Nvidia graphics processing units, according to a person familiar with the matter.
Brave Roman soldiers know the biggest and most modern ship in the fleet has been rammed and will sink. Thanks, Google. Close enough for horse shoes.
I find that anonymous reports and unidentified sources a very good indicator of the effort a news gathering output expends for a story.
Okay, we have some clicky names: OpenAI, Oracle, and Nvidia. Stargate has a weird Captain Video resonance. And Abilene? I think it is known as the official storybook capital of the US. Oh, yes, there are wide open spaces, some water, some power due to Texas’ outstanding power generation set up, and cows.
The article states:
Oracle secured the site, ordered the hardware, and spent billions of dollars on construction and staff, with the expectation of going bigger.
But — and this is an important “but” — the story presents:
OpenAI is no longer planning to expand its partnership with Oracle in Abilene, Texas, home to the Stargate data center, because it wants clusters with newer generations of Nvidia graphics processing units
What is behind this Oracle problem? The write up asserts:
For the companies building frontier models, the smallest improvement in performance could equate to huge gaps in model benchmarks and rankings, which are closely followed by developers and translate directly to usage, revenue, and valuation. That all points to a bigger problem at play. For infrastructure companies, securing a site, connecting power and standing up a facility takes 12 to 24 months at minimum. But customers want the latest and greatest, and they’re tracking the yearly chip upgrades.
The “real” news story seems to be in the last paragraph, and I quote:
Beyond Oracle, GPU depreciation is a risk for the broader market and could have ramifications across the AI landscape. Every infrastructure deal signed today may result in a commitment to outdated hardware before the power is even connected.
The issue is time. Data centers require time to build. Outfits like Nvidia operate on Silicon Valley clock cycles. The challenge of getting the two approaches to time to line up is difficult. When the timing does not line up, multiple costs stack up. One hopes that whiz kids can figure out how to deal with two clocks without collapsing under the fungible weight of the intangible clicks. In a James Bond film, the fictional hero stopped the ticking clock at 007. Without his skills, the bomb housing to which he had been handcuffed would have detonated.
Is James available to provide consulting advice to Oracle- and OpenAI-type outfits?
Stephen E Arnold, March 18, 2026
A Pragmatic Look at Data Center Power: Just Turn Everything on at Once. No Problemo.
March 12, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
A happy honk to the UK online information service the Register. “Your Datacenter’s Power Architecture Called. It’s Not Happy” presents some useful information about how AI compute data centers operate when the “on” button is pressed. The article explores in reasonably a non-EE geeky way some of the different facets of taking a corn field in Louisiana or a tobacco farm in Tennessee and building a five soccer field size data center. Heck, make it bigger. Land is cheaper than in Manhattan. Go for 20 soccer fields of AI compute hardware.
Yep, AI is close enough for horseshoed. Thanks, MidJourney. Nice fire.
The write up says:
GPU clusters and AI accelerators don’t operate on the old rules. They don’t ask for 15 kW. They demand hundreds of kilowatts per rack, an order-of-magnitude leap that legacy electrical and thermal architectures were never designed to survive. The comfortable assumptions baked into decades of datacenter design are now liabilities, and the industry is facing a reckoning it can no longer defer.
What? Are big time AI outfits kicking the can down the road or simply ignoring the fact that “power” is just there. It’s like water. In their accelerated lives, power has not been top of mind. Now the whiz kids at big tech AI companies have an opportunity to learn about the challenges their acres of computers and assorted gizmos will require.
The write up identifies a number of issues; for example:
Let’s talk about the current-squared problem and resistive losses. Because power loss scales with the square of the current, even small reductions in current lead to significant increases in efficiency. The power distribution efficiency is governed by Joule resistive loss (Ploss</loss> = I2R).
If you are not into this power lingo, the author seems to be saying that the new AI compute data centers require some extra engineering. Why? The pricey CPUs and GPUs can run hot. These devices plus RAM and solid state storage are voltage?sensitive. Even idle or under utilized racks can run at higher temperatures than old-fashioned data center racks. When demand hits to create a video for grandma, the AI compute system has to deal with heat. AI centric systems can throttle themselves. Pushing harder means that electrical noise can become a factor. Furthermore, failover is harder because the stacks are non?identical and latency?sensitive.
How do you deal with infrastructure challenges, power, and cooling? Answer: Novel engineering and money. The problem is that these big AI compute data centers are now the equivalent of the room-sized mainframes in the late 1950s. The solutions make 20 somethings laugh when the Smithsonian in DC puts some of its historical hardware on display.
The real equations are not the EE marvels. Nope, the basic equations pivot on the cost of time required to engineer, design, test, and manufacture the components necessary to keep power draw lower and temperatures even lower. At the same time, the solutions have to allow the CPUs and GPUs to go fast.
Accelerationists think about software. Accountants and people with common sense think about plumbing. Experienced professionals think about time and how much it costs to crunch quite challenging engineering into tiny intervals. Speed can kill data centers, the financial dreams of high tech whiz kids, and some businesses.
What happens when one tries to scale these nifty data centers? No big deal. We are big tech doing AI. We have the answers.
Stephen E Arnold, March 12, 2026
Who Pays for Electricity?
March 9, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
The cost of electricity is an interesting topic. Figuring out power demand and the cost of the infrastructure requires research, math, experience, and knowledge of technical options. I worked years ago with a utility rate expert. He was a specialist, and he was, no pun intended, in demand.
I thought about the easy assumption that power is just available. Buy a gizmo, plug it in, and it turns on. The electric bill arrives, and it is relatively predictable. Rates typically don’t gyrate chaotically. When my family lived in Brazil, the power would go off but the “cost” of the electricity was reasonably predictable. What happens when one assumes that a data center can be built and plugged in? The electricity is there and the bill arrives. No big deal.
When the batteries go dead, recharging may be a challenge. Thanks, Venice.ai. Good enough.
Nope. Big deal. Utility demand forecasting is tricky when nothing wobbles too far from the forecasters’ and analysts’ projections. Plugging in numerous data centers means that the complex machinery for power generation has to be refigured. Behind the scenes of what is to most people a power generation company are many moving parts. Out of sight and out of mind does not mean easy to operate and trivial to scale.
I thought about this assumption that electricity is just “there” when I read “Tech Giants Sign Energy Pledge at White House Ahead of Midterms.” The operative word is “pledge.” Promising to pay for electricity required to run the data centers some companies are building, trying to build, planning to build, and hoping to operate is just that a promise. A promise is not cash. Furthermore, I am not sure the “tech giants” know the cost of the electricity their AI factories will require.
The BBC article “Tech Firms Pledge to Pay for AI data Center Power Costs. But Will They?” cuts to the core of the electricity problem. The write up states:
The tech pledge may be difficult to enforce, said John Quigley, a senior fellow at the Kleinman Center for Energy Policy at the University of Pennsylvania. He cited the multiple layers of government, grid managers and electricity regulators involved in power projects.
Saying is one thing. Doing is another.
When it comes to power generation, fast talk may provide superficial reassurance to some people. The shift from a pledge to forking over money may not be the same as putting one’s luxury car in gear and allowing the smart self driving vehicle to whisk the passengers to an off site face to face meeting. How is that working out by the way?
Several observations:
- Rolling out cost effective power generation takes time, specialty manufacturing, and planning. One does not ring up a company and have a modular nuclear generating machine dropped off at the loading dock.
- Creating electricity and getting it where it needs to be are shotgun marriages of old-fashioned technology, advanced materials, and infrastructure both physical and digital. The costs to ensure a happy, stable union are usually steep and difficult to project with accuracy.
- People living near a new or expanded power generation facility, infrastructure, or a big data center might throw a wrench in the works. What few people realize is that power generation is one of the more vulnerable “it’s just there” services. Angry consumers can create some tricky challenges easily and quickly.
Net net: Pledges are not power. Pledges are words. Power generation makes modern life possible, yet few whiz kids know anything about the business. When the power goes away as it did where my family lived in Brazil, one needs a Plan B. A pledge is not a Plan B. A pledge is public relations.
Stephen E Arnold, March 9, 2026
Google: Cracks in the Facade of Smart
June 2, 2019
I find it amusing that a company with the smartest people in the world cannot fail gracefully. When the GOOG goes down hard. I discovered this chugging along from rural Kentucky to a rural location in South Carolina. Google did not deliver. Once I was able to fire up a connection which actually worked, I read “Google Outage Takes Down YouTube, Gmail, and Snapchat in Parts of US.” I learned:
Discord, Snapchat, and Vimeo users are also experiencing issues logging into the apps, and these all use Google Cloud on the backend. “We are experiencing high levels of network congestion in the eastern USA, affecting multiple services in Google Cloud, GSuite and YouTube,” says a Google spokesperson in a statement to The Verge. “Users may see slow performance or intermittent errors. We believe we have identified the root cause of the congestion and expect to return to normal service shortly.”
Now about those smart people. Are too many trying to abandon assignments which have zero future for the zippier work? Of course not. Google does not have Android fragmentation or other technical weaknesses. I would suggest that some work needs to be done on foundational services.
Stephen E Arnold, June 2, 2019
Apple and Google: Beacons Beckon and Inform
May 17, 2018
I read “Apple News Officially Lets Publishers Use Google’s DoubleClick to Serve Ads.” The main idea is that a couple of large companies are teaming up in a way that matters: Advertising or at least keeping some folks happy. I don’t have a dog in this fight, but the news, if accurate, reminded me that making money is a goal which makes bad experiences like appointing some individuals to a company’s Board of Directors go away.
The write up asserts:
The ad-targeting options break down into two categories: context-based targeting and audience-based targeting. Contextual ads can be aimed based on the article’s publisher, its content category within Apple News and the tags a publisher appends to the individual article as well as according to whether it appears on an iPhone or iPad. Audience-wise, the ads can be targeted by a person’s location (though only at the designated market area level), their gender and their age group.
What’s this mean? A partial answer can be excavated from the glorious prose of these public documents:
- US7039599, Method and apparatus for automatic placement of advertising
- US7085682, System and method for analyzing website activity
- US7349827, System and method for reporting website activity based on inferred attribution methodology
- US7844488, Method of delivery, targeting, and measuring advertising over networks.
These documents can be downloaded from the USPTO. Think of the link as a beacon which is a heck of a lot less effective than the inventions disclosed in these patents.
Stephen E Arnold, May 17, 2018
Ebay Is Sold Structured Data
January 19, 2016
PayPal and eBay split in 2015 and many people thought it was a poor mistake on eBay’s part. However, eBay has recouped any potential loses by record profits and more than 159 million sellers. Channel Advisor explains that one of the reasons eBay has grown so much is due to its incorporation of structured data and its importance for organic search in the article, “eBay Moves Towards Structured Data-And Why It Matters.”
As an avid eBay buyer and seller, I have been impressed with the new changes in eBay’s demand for structured data. In the past, if you wanted to find anything on eBay you had to go directly to the Web site and dig through results. Sometimes you could find results on Google or another search engine, but these were usually cached auctions. Since the switchover, eBay listings are prominent within Google’s search results. What is even better is how accurate they are!
EBay has turned to structured data as a way to compete with Amazon. While this is beneficial in the long run, it forces sellers to refocus their strategies. The article gives some great tips on how to improve your listings for the best organic search effectiveness. What eBay is demanding now is item specifics so items are placed in the right categories and also helps buyers make more informed decisions. Product identifiers are now very important and mandatory in many categories. These include item specifics such as UPCs, ISBNs, MPNs, GTINs, and more. The goal with all this extra information is to increase visibility in Google and eBay search results.
“In addition to the above benefits, adding identifiers will give you:
- The ability to match your item with a product from the more robust eBay catalogue
- More accurate pricing guidance when you list your items
- Trending price alerts — when your listings are priced lower than the trending price
EBay suggests adding identifiers even if they’re not yet required for your category – doing so will earn you an early competitive edge.”
EBay used to be the one-stop shopping destination online, but Amazon has quickly stolen that title from them. With more detailed listings and visibility in Google, eBay is sure to win back customers.
Whitney Grace, January 19, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Perceptive Software Builds Offices with Emphasis on Group Spaces
January 29, 2015
The article titled Johnson County’s Coolest Offices Are Places Where People Want to Work on The Kansas City Star praises Perceptive Software for its interest in architecture and search. Their new office building boasts wide-open spaces and flexible work areas. There are still cubicles, but the group spaces are more thoughtfully planned to be inviting and inspiring. The article states,
“We wanted the building to be a visual depiction of our fun, collaborative culture,” said Megan McClendon, spokeswoman for Perceptive Software, as she began a tour of the new 238,000-square-foot building in its central atrium. “We have flexible work options for employees to work at home or wherever, but employees are coming in to work more often. They don’t have to be here, but they do.”
Small touches can make a big difference in environment. Items like the V-shaped table for videoconferencing are the round table of corporate equality, enabling everyone to see and be seen. Having a dry cleaners in the building and a gym make errands and personal time more accessible and convenient, improving morale and even allowing for employees to spend more time at work. 700 people work in the new offices, and perhaps the most revolutionary aspect of their construction is the attempt to dismantle the “corner office with a view.” Instead, it is the group spaces that get the windows, while the private officers are in the middle. Clearly, this sort of innovation is exciting in the value it sets on employee spaces.
Chelsea Kerwin, January 29, 2014
Sponsored by ArnoldIT.com, developer of Augmentext
-
- Subscribe to Beyond Search
Feature archive
News archive
-
