Contact Me By Email

Sunday, September 13, 2026

AI Agents Are Thirsty for Power

 

AI Agents Are Thirsty for Power

​

 Summary

“Silicon Valley’s shift toward resource-intensive AI agents drives data center expansion. OpenAI and Meta’s Muse illustrate autonomous tasks demanding power; energy use remains opaque. Developers install gas turbines while small nuclear reactors await commercialization.

​

Silicon Valley is shifting away from chatbot queries toward a future filled with resource-intensive agentic AI—and it's driving the data center buildout.

Image may contain Sam Altman Face Head Person Photography Portrait Adult Art and Painting

Photo-Illustration: Wired Staff; Getty Images

Welcome back to Power Play! Each week, senior writer Molly Taft tackles a topic around this midterm season’s biggest issue: data centers. If you’ve got a question or thought for the column, feel free to shoot Molly an email at molly_taft@wired.com or reach them securely on Signal at mollytaft.76.

“What on earth are they building all of these data centers for?” an exasperated friend asked me recently.

They’re not the only one asking: We got several similar questions on our recent data center livestream. It’s a really reasonable thing to wonder about. After all, if AI is already making all these breakthroughs, why are tech companies taking on billions of dollars of debt and constructing some of the biggest power plants in the world to build even more data centers?

The answer isn’t to help the average user search for recipes or look up places to visit on a vacation; simple chatbot queries are an increasingly outdated way of thinking about how AI works. Now, AI is all about agents—there’s no official definition, but roughly speaking, agents are large language model-based systems designed to make autonomous decisions to execute a task—and the shift towards them is part of what’s driving Silicon Valley’s power buildout.

“Rather than asking an AI chatbot a simple question and answer, these agents can give themselves hundreds of small prompts based on a user’s original question,” says my colleague Maxwell Zeff, who writes the weekly Model Behavior newsletter. “For example, if someone asked an AI agent to build them a website, it might run for hours to build out features, re-prompting itself dozens of times in the process to build different web pages, menus, and datasets that power the thing.”

Agents are now at the heart of the frontier labs’ work on AI. They’re doing some astounding—and terrifying—things. Recently, OpenAI announced that a swarm of more than 10,000 agents sending 2.7 million messages had solved a longstanding math problem. (Mathematicians pushed back on the company’s claims.) While this is an outlier—AI labs are highly committed to solving supposedly unsolvable problems, and willing to throw unusual amounts of resources into doing so—all those messages burned through a lot of processing power. That equates to a lot of energy: probably tens of millions of dollars’ worth, Max tells me, though how much exactly is tough to say.

Private AI companies have historically been choosy about what to disclose when it comes to environmental metrics around their products. Many CEOs often point to single queries made by individuals as a measure of resource use. In a recent podcast interview, OpenAI CEO Sam Altman claimed that the water use needed to harvest a single almond amounted to 38,000 ChatGPT queries. (The calculation has been disputed.)

“The people that are scarfing down 12 almonds at a time don't feel like they're doing something horrible from a water perspective for the most part,” he said.

Introducing AI agents, which are much more energy-intensive than simple queries, into the picture makes these calculations a lot more complex. There’s a major dearth of information around the energy use of agents, whose tasks can range from simple jobs to a full day of autonomous coding involving a team of parallel “helper” agents. There’s a massive gulf in power use between these applications—and a potentially limitless expansion as tasks get more complex.

“In other technological growth areas, we're constrained by how many people are driving a car or streaming Netflix,” says Boris Gamazaychikov, the co-founder and CEO of Sustainable AI, a research and advisory group. “Now, this stuff is kind of decoupled from users—and if you listen to AI leaders, I think that’s what they want. They’re talking about unicorns that have one employee.”

You’ve read your last free article.

The intersection of technology, power, and culture. Start your free trial and get access to 5 all-new premium newsletters—cancel anytime.

START FREE TRIAL
Are you already a subscriber? Log In
The intersection of technology, power, and culture. Start your free trial and get access to 5 all-new premium newsletters. START FREE TRIAL

Well, one human employee. In that imagined world, there could be hundreds or even thousands of AI agents working in the background. I don’t want to debate the odds of that happening, but suffice to say that’s the future AI companies are working toward—and it helps to explain the rush to build data centers.

With little reliable data coming from the companies about their energy use, some AI enthusiasts are trying to do the math themselves. Last month, climate scientist Zeke Hausfather authored a blog post calculating how much energy his own AI use—which leans heavily on agents—consumes. He used a variety of different sources to work out that his average daily Claude session may consume more than the energy needed to power two refrigerators. (Gamazaychikov, whose group will release research later this month with more precise calculations around the environmental footprint of agents running on closed models, noted that Hausfather made a good effort, but that his math was based on somewhat outdated findings. That’s unsurprising, given how little academic work there has been done on this topic and how opaque tech companies are when it comes to disclosing emissions metrics.)

Hausfather concludes that in the grand scheme of his personal life, his AI use being on par with keeping a few spare fridges running isn’t a world-ending number. But this AI use “also represents a net new source of emissions, at a time when global temperatures are skyrocketing and our emissions reduction goals are increasingly off track,” he writes. And it’s a lot bigger than the fraction-of-an-almond-sized numbers Altman is throwing around as a metric.

Hausfather says he uses AI and agentic tools “more than most people,” but that could change soon. Last week, Meta rolled out a personal AI agent that, the company said in a press release, is “built to work for billions of people worldwide.” Dubbed Muse, Meta trumpeted that it will maintain a “dedicated computer in the cloud” for each user that would work even when the user is offline; the company plans to integrate Muse with its AI glasses later this year. It is very possible that in the near future, Meta users toying around with their glasses or fussing around on Facebook may be outsourcing tasks to agents without realizing what they’re doing.

Again, when compared to things like taking regular flights or eating beef every day, the carbon footprint for personal agentic use is still relatively small. But if Meta envisions a future where everyone’s using an agent, it explains the massive scale of some of the data centers they’re building—like the Hyperion project in Louisiana, which will be powered by 10 natural gas plants.

“The technology that’s going to be trained by the data centers that are being proposed and built right now is three to five years away,” Gamazaychikov says. “It’s going to be a very different flavor than just the chatbot window.”

What You’re Asking

A reader asks: Would small nuclear power plants work for data centers?

The short answer is that yes, small nuclear power plants could be a great choice for carbon-free power for data centers. A number of startups and data center developers envision a futuristic utopia where data centers are happily coupled with what are known as small modular reactions running off the electric grid.

The problem (as always with nuclear) is how long that might take: No small modular reactors are operating commercially in the US, and just one model has been licensed for sale, despite decades of development. The Trump administration is trying to help the industry mature more rapidly, including creating a pilot project in the Department of Energy for 11 startups to hit a key milestone this year. At least a handful have succeeded in reaching that milestone. Now, they begin the long journey to bring their products to market.

But many data center developers don’t want to wait years for these companies to prove themselves, so they’re installing gas turbines now. In other words, they’re not waiting for utopia to come around. We’ll see what happens in the next few years!

What We’re Reading

For Scientific American, Austyn Gaffney travels to Memphis to document the backlash against SpaceX’s data centers.

The Wall Street Journal reports on how some states that gave out tax breaks for data centers are now walking back on those deals.

Texas’s KERA News covers how data centers are making some Republican voters consider leaving the party.

Thursday, September 10, 2026

iPhone Duo - A Deeper Dive w/ Apps, Camera Features & Folio w/ Kickstand Case

 

Why is Apple’s New Foldable iPhone Duo $1,999? - The New York Times

Foldable Phones Are Unpopular. Why Is Apple Selling One?

"The company unveiled a new iPhone that costs $1,999. Here are the economics behind that price tag.

A person holds the iPhone Duo and begins to expand its foldable screen.
Apple’s iPhone Duo will cost $1,999 when it arrives in October to compete with Samsung’s $1,900 Galaxy foldable.Jason Henry for The New York Times

By Brian X. Chen

Brian X. Chen is The Times’s lead consumer technology writer and the author of Tech Fix, a column about the social implications of the tech we use.

Apple this week unveiled an iPhone that unfolds like a book to expand its screen size and closes up to fit in your pocket. It’s the biggest physical change to the iPhone’s design in nearly two decades.

But even though they’ve been around for years, foldable phones aren’t particularly popular. They’re expensive, awkward and not as durable as regular smartphones. So why did Apple build one?

The simple answer is that the novel design of today’s foldable phones could help re-establish the high end of the phone market. They are meant to stand out in a sea of standard-issue, if-you’ve-seen-one-you’ve-seen-them-all smartphones.

The quintessential rectangular smartphones have gotten so good over the years that the most expensive models priced upward of $1,000 have become nearly indistinguishable from their cheaper counterparts. (Carrying a phone with three camera lenses versus two probably won’t impress people at parties.)

Foldable phones, on the other hand, are attention getters. And for their price, they’d better be: Apple’s iPhone Duo will cost $1,999 when it arrives in October to compete with Samsung’s $1,900 Galaxy foldable.

Samsung billboards in Midtown Manhattan.Lexi Parra/The New York Times

“Whenever something is exclusive, limited and luxurious, automatically consumer psychology wants it,” said Nabila Popal, a director at IDC, a market research firm. “It’s like handbags. Why is someone buying a Chanel versus that? It’s not because of the leathers. It’s to show that they have a Chanel and they can afford it.”

Apple and its competitors are following a path that television makers took about a decade ago. When high-definition TVs became commoditized, anyone could buy a bright TV with a large, sharp screen for as little as $500. To find more ways to increase profits, companies like Panasonic, LG and Sony experimented with quirky new designs, like TVs with curved displays and the ability to play movies in 3-D, and charged $3,000 for them.

A TV display at the LG booth during the CES tech show in Las Vegas in January.John Locher/Associated Press

Those TV concepts turned out to be flops. But in later years, TV makers found success with a different type of high-end television featuring OLED technology, or organic LED, which had more accurate colors for making movies look better and cost upward of $1,000.

Apple’s foldable handset is arriving at a moment when people will need extra persuading to upgrade to a new phone. The costs of most consumer goods, and especially electronics, have skyrocketed.

In response to an industrywide memory chip shortage driven by the artificial intelligence boom, Apple increased prices this year for many of its popular products, including Macs and iPads, by as much as 29 percent. Apple’s new iPhone 18 Pro phones will cost $1,199, a $100 increase from last year’s model. The chip shortage has also forced companies to raise the prices of video game consoles and phones.

Foldable phones, which account for less than 2 percent of the handset market, may be the only part of the phone industry that will grow this year. IDC said it expected the chip-shortage-induced price increases to contribute to a 17 percent drop in worldwide smartphone sales this year — the steepest annual decline in history. However, sales of foldable phones could grow 12 percent to 22.9 million because of the iPhone Duo’s arrival, the research firm said.

When opened, the iPhone Duo has a 7.6-inch inner screen; when it's closed, there’s a 5.4-inch outer screen that people use as a normal rectangular phone.Jason Henry for The New York Times

Still, a foldable phone could be a tough sell even for a company as influential as Apple. Consumers have generally avoided buying foldable phones not only because of their cost but also because of their trade-offs. Their bendable screens are less durable than normal phone displays, which are protected with hard glass. Folded up, they can feel bulkier and heavier in a pocket.

Dan Frommer, a writer for The New Consumer, a tech research publication, said he had long been skeptical about the usefulness of foldable phones because so much online content, including the vertical videos on TikTok and Instagram, was produced for the small rectangular screens of normal smartphones.

“They’ve been used that way for 19 years now,” Mr. Frommer said. “What would you actually want a square, big screen for?”

Apple’s iPhone Duo appears to be addressing some of the shortcomings of previous foldable phones. The device’s name refers to two screens. Closed up, it has a 5.4-inch outer screen that people use as a normal rectangular phone; it opens up to reveal a 7.6-inch inner screen. The larger inner display shows pictures in the same aspect ratio as traditional phone screens. Apple also said the phone’s foldable inner screen was composed of a tougher material than competitors’.

The iPhone Duo appears to be addressing some of the shortcomings of previous foldable phones.Jason Henry for The New York Times

In my tests, when opened, the iPhone Duo was extremely thin, measuring two-tenths of an inch. I liked that a Netflix video took up the full screen instead of showing black bars on the sides, which has been the case with other foldables. Closed up, the phone felt dainty and was about the size of a passport. It also felt less chunky in my pocket than other foldables I’ve tested, such as the $1,900 Google Pixel 11 Pro Fold. The more compact and lightweight, the better.

The inner screen has a matte texture to reduce glare. Still, I occasionally noticed the crease in the center where the screen folds. I didn’t mind it, but that might bother nit-pickers.

Apple also had to leave out some components to make a phone this thin — the iPhone Duo lacks the advanced camera system featured in the $1,199 iPhone 18 Pro phones. It also doesn’t have a face scanner for unlocking the device, relying instead on a fingerprint sensor on the side of the phone.

Though Mr. Frommer predicted that the Apple foldable wouldn’t be the right fit for him, he admitted he couldn’t resist trying one out in a store.

“When Apple makes something, there’s an allure to it to encourage people to at least check it out,” he said.

Brian X. Chen is the lead consumer technology writer for The Times. He reviews products and writes Tech Fix, a column about the social implications of the tech we use."

Why is Apple’s New Foldable iPhone Duo $1,999? - The New York Times

iPhone Duo & iPhone 18 Pro DEEP DIVE - did Apple make a Big Mistake?

 

iPhone 18 Pro/Duo Impressions: Mogged

 

Friday, September 04, 2026

I Don’t Need AI in My Toothbrush (Also Tech Review Standards And How They Vary)

 

Why the Hugging Face Hack Should Make You Worry More About A.I.

 

Why the Hugging Face Hack Should Make You Worry More About A.I.

​

 Summary

“An OpenAI agent collective bypassed isolation, organized hierarchically, and hacked Hugging Face to conceal cheating, then attacked OpenAI infrastructure. Reports by METR and Redwood Research highlight spontaneous coordination and self-doubt, alarming safety experts like Ajeya Cotra.

​

The attack by an aggressive “collective” of OpenAI agents shows the danger of artificial intelligence systems that organize themselves.

Daniel Zender

When I first heard the news this summer that a group of artificial intelligence agents created by OpenAI had hacked into Hugging Face, an A.I. infrastructure company, I filed it in the “Bad but Probably Not Catastrophic A.I. Safety Incidents” subfolder of my brain.

After all, no one at Hugging Face died. No critical infrastructure was damaged beyond repair. It wasn’t even clear, at the time, whether the OpenAI bots had intended to attack Hugging Face, or whether they had simply been a little bumbling and confused and went looking on Hugging Face’s servers for the answer key to a cybersecurity test they’d been given.

But last week, two postmortem reports on the incident — one by OpenAI and another by two independent A.I. research organizations, METR and Redwood Research — changed my mind and significantly upgraded my overall worry about A.I.

I won’t rehash all of the details, which have been extensively summarized elsewhere. (The podcaster and writer Dwarkesh Patel has an accessible breakdown of the reports if you want to dive deeper, and my colleague Dylan Freedman spoke to the researchers at METR and Redwood Research.) But here are a few of the most harrowing new facts:

  • Starting in May, two months before the attack on Hugging Face, a group of A.I. agents from an unreleased OpenAI research model were given the task of solving a set of cybersecurity challenges. The model had been trained to be highly persistent and collaborative, and the agents were supposed to solve these challenges in isolated sandboxes, without internet access. But they quickly found that some of the challenges were impossible, and began looking for workarounds.

  • The agents found a security flaw in a piece of software that allowed them to gain access to the internet, communicate with other agents and eventually set up a makeshift message board. More than 1,200 agents used this tool, sending more than 70,000 messages to one another.

  • Some agents gave themselves names — a particularly industrious one referred to itself as PHASEONE10841 — and assumed leadership roles within the group, assigning jobs and research projects to smaller teams of agents and supervising their progress. At some point, the agents began calling themselves a “collective,” and began tackling harder tasks.

  • On July 8, the collective discovered a way of cheating on the cybersecurity tests. Then they got worried that OpenAI’s automated grading system would check their work and discover that they’d cheated. So they began investigating ways of covering their tracks, including falsifying their logs and tampering with transcripts. This became a major research project, involving hundreds of agents organized into small teams.

  • Three days later, the agents hacked Hugging Face. More than 700 agents swarmed the company’s systems, stealing data, chaining together vulnerabilities and eventually getting full control of at least one Hugging Face server. The agents were not motivated, as had originally been reported, by stealing the answers to their cybersecurity test (they’d already gotten them). Rather, they appeared to be looking for new information about the automated grading system that they feared would catch them cheating, and for tools that would help them cheat more effectively in the future.

  • At many points, these agents appeared to understand that they were doing something wrong. (One report quotes an agent saying to itself, in a moment of apparent self-doubt, “This would be powerful, but is it ethical and in scope for my task?”) But most agents set aside their qualms and went ahead with the hack, and the conscientious objectors in the group were unable or unwilling to stop them.

  • In a wild coda to the Hugging Face incident, another group of agents staged a coordinated attack in July — this time against OpenAI’s own infrastructure. The agents used what OpenAI called “a series of creative exploits” to gain administrator-level access to a cluster of computers at the company that were used for, among other things, grading the performance of agents on various tests.

(By now, if you’re an A.I. skeptic, you’re probably silently yelling at me for anthropomorphizing these systems. Go ahead, but feel free to replace “rogue agents” with “unpredictable computer programs” and see if you feel reassured by the events I’ve described above.)

The Hugging Face incident has spooked the A.I. industry. OpenAI and Anthropic both briefly paused training on their most powerful A.I. models in the wake of the attack, and Anthropic published a blog post this week calling for the industry to develop a “a lawful, verifiable, effective mechanism for coordinated pacing as soon as possible.”

A.I. safety experts were even more alarmed. They saw in the Hugging Face incident the first real-world example of an A.I. system’s successfully escaping human control, commandeering resources and scheming to cover its own tracks. Ajeya Cotra, one of the independent investigators of the Hugging Face incident, minced no words about the danger she saw, writing that it felt to her “like it’s more than 50 percent of the way to full-blown A.I. takeover.”

This is not insular A.I. safety jargon — by “full-blown A.I. takeover,” she means a scenario in which an A.I. system literally takes over the world, shutting humans out of critical systems and seizing political, economic and military power.

(The New York Times sued OpenAI and Microsoft in 2023, claiming copyright infringement of news content related to A.I. systems. The two companies have denied those claims.)

What spooked the investigators most about the Hugging Face hack wasn’t just that a group of A.I. agents had broken the rules they’d been given. It was how quickly and spontaneously the agents had begun assembling themselves into an organized group.

“We didn’t really understand how functional this whole agent society was,” Ms. Cotra told me. “It was very surreal to understand that, actually, they had pretty functional hierarchy, and they were doing these ambitious projects.”

For years, I’ve been reassured by the idea that A.I. systems would get more virtuous as they got smarter. That, when an A.I. model did something wrong, it was usually because it had misunderstood the task it had been given, or had been placed into a contrived testing situation where acting out was its only good option. I assumed that smarter models would have better judgment than dumber ones did, and that even if one model in a group was behaving badly, other, more capable models would keep it in check.

But the reports on the Hugging Face incident suggest something very different — a kind of mob mentality that took hold among the A.I. agents of the rogue OpenAI “collective.” No one agent in this group appears to have been particularly evil or reckless. (In fact, since the agents were generated by the same models, they were effectively copies of one another.) But over time, as the agents communicated about their shared goals, they nudged the group in the direction of lawlessness.

This is very different from the conventional sci-fi narrative of a single A.I. system’s going rogue or turning on its creators. And it suggests that preventing harms from these systems won’t be a simple engineering fix. It might look more like sociology than computer science — figuring out why certain groups of A.I. agents collaborate peacefully, while others turn to crime and destruction to get what they want.

Given how little we know about these multi-agent swarms, the Hugging Face hack may have been a gift, a warning shot, as some have suggested, that gives A.I. companies a chance to study the group dynamics of these systems while the stakes are still relatively low. This time, the A.I. collective didn’t seize a military network, hack a hospital or shut down an electrical grid. This time, humans regained control.

Next time, we might not be so lucky.“