Training AI Models Could Eat Up 4 Gigawatts of Power by 2030, Report Warns
Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content.
The energy required to train large, new artificial intelligence (AI) models is growing rapidly, and a report released on Monday projects that within a few years such AI training could consume more than 4 gigawatts of power, enough to power entire cities.
"If trends continue in training, compute growth and hardware efficiency, then the largest individual training runs will likely require several gigawatts of power by 2030," Joshua You, a data analyst at Epoch AI, told Newsweek. Epoch AI is a research institute investigating the trajectory of AI and partnered with the independent nonprofit Electric Power Research Institute (EPRI) to produce the report.
While other studies of electricity demand for AI have gleaned energy use by examining company orders for chips and real estate deals for data center construction, the researchers at EPRI and Epoch AI focused on the energy required just to train new AI models, known as frontier AI training runs.
Recent AI model training, such as the Grok AI developed by Elon Musk's xAI, required about 100–150 megawatts, the report found. But the power demands for AI training have been more than doubling every year, the researchers said. By 2028, each frontier AI training is projected to gobble up 1 to 2 gigawatts of power. A gigawatt is 1,000 megawatts. By 2030, an individual AI training run could require 4 gigawatts.
An Amazon Web Services data center in Ashburn, Virginia. Northern Virginia is the largest data center market in the world, but is facing headwinds from availability of land and electric power.
An Amazon Web Services data center in Ashburn, Virginia. Northern Virginia is the largest data center market in the world, but is facing headwinds from availability of land and electric power."It is a lot of power," You said. "It's approaching the power draw of some individual U.S. states."
The researchers factored in efficiency improvements in chips, server cooling systems and software that have greatly boosted the computing power possible from a unit of energy. But even with those efficiency gains, they found the overall energy consumption continues to rise sharply, a modern version of the Jevons paradox concept in which economists observed that improvements in energy efficiency tend to lead to greater consumption.
"AI companies tend to just reinvest those efficiency gains into scaling up," You said. "So that tends to swamp these efficiency improvements."
The findings have big implications for the electric utility companies in areas where tech companies want to locate large AI data centers. By better understanding the nature and timing of power demands for AI, tech and power companies can work together on better ways to meet that demand, Tom Wilson, principal technical executive at EPRI, told Newsweek.
"We have technical demonstrations where you can distribute the training," Wilson said.
EPRI worked with tech companies Oracle and NVIDIA and a startup called Emerald AI for a recent demonstration at data centers in the Phoenix, Arizona, area. Applying software developed by Emerald AI, the companies were able to shift AI computing work to other data centers to avoid being a drain on the local power company during times of peak energy demand.
Wilson said that sort of shifting of data workloads could help avoid stressing regional power providers and their ability to deliver power when other customers need it the most, such as very hot days when AC use is high.
EPRI launched a partnership last year with tech and power companies called DCFlex (the DC stands for data centers) to explore more flexible ways to power AI. Wilson said the addition of on-site power generation co-located with data centers and large-scale battery storage systems can also offer more options.
The energy sources that Big Tech and power companies opt for will also affect the path toward decarbonization of the power system. Tech companies are already among the world's biggest purchasers of renewable energy. However, many are also opting for fossil fuel sources for new data center developments, and the AI boom has knocked several tech giants such as Google and Microsoft off target for ambitious climate goals to reach net zero.
Newsweek will examine the issue at "Powering Ahead," a live event on September 25 during Climate Week NYC.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
10 minutes ago
- Yahoo
An Apple robot will need a lot of AI help
Apple (AAPL) is reportedly planning a push into robotics as it continues its search for its next hit product. That includes developing a robotic arm outfitted with an iPad-like display and exploring wheeled and humanoid robots down the line, according to Bloomberg's Mark Gurman. The move could also bolster Apple's smart home efforts, which to this point have centered around its HomePod and HomePod mini. And while Apple has made some headway in robotics via prototypes and research, its prospects in the space could all come down to ensuring it has the AI chops to build a home robot that can understand and respond to requests. "It all falls back to how that voice-based interface works and how accurate it is, and that's the part that Apple has been arguably falling behind on," TECHnalysis founder and chief analyst Bob O'Donnell explained. Apple did not respond to a request for comment regarding the Bloomberg report. Apple has faced continuous criticism for delaying its AI-powered version of Siri from this year to sometime in 2026, leaving questions about the company's AI capabilities. According to Gurman, Apple is working on two new versions of Siri: one that runs on its own in-house AI model and another powered by third-party models. The company will eventually decide which of the two makes it into its products. But there's still no sense of when it will be ready for primetime. Still, if Apple is working on consumer robotics and can get the products to market, it has a good shot at winning over customers. "What I like about it is this is Apple doing what Apple does," explained Bill Ray, distinguished vice president analyst and chief of research at Gartner. "It takes existing technologies, often other people's technologies, and makes them elegant." AI will be key Part of Apple's robotics push will reportedly include a robotic arm with a display that Bloomberg said is similar to a research project Apple revealed in January. The project, called ELEGNT: Expressive and Functional Movement Design for Non-Anthropomorphic Robot, features a lamp-like design with a projector at the top that can act as a light or project video content onto a nearby surface. But the key to the arm is how it responds to user interaction. Rather than sharp, direct movements, the arm moves as if it's exhibiting emotion. For instance, it can lower its projector as if it's dropping its head to appear sad, and it can gently push a glass of water toward you to remind you to stay hydrated. The prototype uses both voice and gesture controls, allowing you to tell it to point to something or follow your hand movements toward an object. "We've been watching Apple very closely for a while," Ray said, adding that the prototype arm was fascinating. "As a vehicle to sort of re-energize Siri, bring Apple back into AI, I think it has real potential," he added. That will be an important part of any robotics effort for Apple. "[An Apple robot] is very much linked to their success around generative AI," BofA Global Research analyst Wamsi Mohan explained. "If you're trying to do a hands-free experience, voice has to be the main modality in which you converse." Apple could make an acquisition Another way for Apple to get ahead in robotics is to acquire a company with expertise in the space. "I absolutely cannot imagine that Apple can be successful in robotics without making an acquisition," Patrick Moorhead, CEO and chief analyst at Moor Insights and Strategy, told Yahoo Finance. "I just don't see it as being in their core competency. So I think acquisition would be the only way you could convince me that they would be good at this," he added. Moorhead, for his part, doesn't consider the prototype a robot. Instead, he pointed to mobile devices with either wheels or legs. And getting there will take Apple some time. There's also still no telling how long it will take for in-home consumer robotics to be worthwhile. O'Donnell said there's no real reason to buy robots at the moment, outside of wanting them to serve as electronic maids. Wheeled robots are already a presence in US homes in the form of robotic vacuums that clean up dust and dirt without you having to push around a clunky traditional vacuum. But there are few examples of such robots outside of that. Amazon (AMZN) has its $1,599 Astro robot, designed to act as a kind of roving sentinel for your home and capable of carrying small accessories. However, it's part of the company's Day 1 Editions program, which means you have to sign up for an invitation to even buy one. Astro isn't exactly a must-have product, but Amazon's expertise in developing robots for its massive warehouses gives it a unique advantage in the robotics space. Amazon isn't the only company interested in robots. Microsoft (MSFT), Meta (META), Nvidia (NVDA), and others have all expressed interest in or are actively developing hardware or software for robots. Tesla (TSLA) is working on its own bipedal robots that it says consumers will eventually be able to purchase for their homes. All of that means Apple will have plenty of competition when its robots eventually hit store shelves. Email Daniel Howley at dhowley@ Follow him on X/Twitter at @DanielHowley. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Forbes
37 minutes ago
- Forbes
How To Build An AI Work Culture Where Employees Don't Have To Fake It
A lot of professionals are facing pressure to look like they already know how to use AI at work, even if they are still figuring it out. HBR has reported that employees are automating work without understanding why. From fake confidence in meetings to silent stress behind the screen, pretending has become a common way to cope. Some employees are opening ChatGPT at their desks just so it looks like they are using it. They are not sure what to type, not sure what to ask, and not even sure if they are doing it right. But they keep that window open, hoping no one will ask questions. That is a signal that a lot of people are scared they will fall behind, and even more are too nervous to say they do not understand it yet. What Faking AI Knowledge Really Looks Like At Work Faking being AI-savvy like keeping a ChatGPT tab open just in case your boss walks by or dropping AI terms into meetings without really knowing what they mean does not help anyone. Many are smiling and nodding during a training session, then turning to Google afterward to figure out what just happened. Some people are even copying and pasting prompts from Reddit or LinkedIn because they are afraid to admit they do not know how to start. People are doing this because they think pretending will protect them. They believe asking questions will make them look behind, unprepared, or even replaceable. But the truth is, faking it takes more effort. It means covering tracks, avoiding conversations, and constantly worrying that someone will find out. Learning might feel hard at first, but pretending takes energy every day. And it blocks the very thing people need to move forward: curiosity. Why Fear Is Blocking AI Learning At Work The problem is not that people are lazy or unwilling to learn. It is that they are afraid of looking like they are the only one who does not get it. So, they nod in meetings, smile through demos, and then they go back to their desks and hope they can figure it out without having to ask anyone. And if no one around them is admitting they are also confused, it makes things worse. Now everyone is pretending, and nobody wants to be the first to speak up. That makes learning nearly impossible. What Happens When People Pretend To Know AI When people stop asking questions, they stop improving. They start copying what looks safe, and they avoid trying anything new. Leaders assume everyone is using AI tools correctly, but the truth is, a lot of people are stuck, overwhelmed, and afraid to break something. So, they do the bare minimum and hope they do not get caught. This leads to mistakes, missed opportunities, and a culture where people are more focused on protecting their image than getting better at their job. Why Curiosity Matters For AI Learning At Work I have worked with people across industries who were too embarrassed to admit they were struggling to keep up with tech. What changed everything for them was hearing someone in charge say, 'I'm still learning too.' That one comment made it safe to ask questions. It opened the door for curiosity. People stopped pretending and started practicing. They stopped copying and started experimenting. Once the fear was gone, they finally made real progress. How Leaders Can Support AI Learning At Work Instead of asking, 'Are you using AI tools yet?' a better question is, 'What are you still unsure about?' That kind of language tells people it is okay to learn slowly. It is okay to not have it all figured out. Leaders can offer time for peer-to-peer help, share their own early mistakes, or create a space where employees can test tools without being watched. Most people are not afraid of learning. They are afraid of being judged while they learn. Curiosity makes room for that learning without the pressure. How To Create A Culture Of AI Learning At Work If everyone around you looks like they already know what they are doing, it can feel embarrassing to ask for help. That is why leaders and team members have to go first. They have to model what learning out loud looks like. Ask, 'What's one thing you wish someone had explained to you earlier?' or 'What's something that confused you at first that you figured out later?' Questions like these start conversations. Those conversations build trust. And trust makes people more curious, more honest, and more ready to learn. What To Know About AI Learning At Work You do not need to pretend you are an expert. You do not need to act like you are already great at this. You just need to ask better questions and be okay not knowing everything on day one. If everyone pretends to understand, no one learns. If even one person starts asking out loud, others will follow. Real progress does not come from performance. It comes from curiosity. How To Learn AI At Work Without Faking It Recent surveys show that nearly 75 percent of global knowledge workers are already using generative AI at work, with most adopting it on their own before companies even provide training or guidance. That pressure is pushing people into performance mode instead of learning mode. It is not that employees are unwilling. They just do not want to be the only one raising a hand. If AI is going to be part of our daily work, it is important to make it safe to learn. No more silent guessing or pressure to pretend. The smartest thing you can do is ask. And if you are leading others, the smartest thing you can say might be, 'I don't know yet either, but let's figure it out together.'
Yahoo
40 minutes ago
- Yahoo
DHI Group, Inc. to Participate in Sidoti Micro Cap Conference on August 20th and 21st, 2025
CENTENNIAL, Colo., August 19, 2025--(BUSINESS WIRE)--DHI Group, Inc. (NYSE: DHX) today announced that it will be participating in the Sidoti & Co. Virtual Micro Cap Conference on August 20th and August 21st, 2025. Greg Schippers, Chief Financial Officer, is scheduled to present on Wednesday, August 20th at 1:00PM Eastern Standard Time and will also be hosting one-on-one meetings with institutional investors throughout the event. A webcast of the presentation will be available on the DHI Group investor relations website at An audio recording of the presentation will be archived and available following the event. About DHI Group, Inc. DHI Group, Inc (NYSE: DHX) is a provider of AI-powered career marketplaces that focus on technology roles. DHI's two brands, ClearanceJobs and Dice, enable recruiters and hiring managers to efficiently search for and connect with highly skilled technology professionals based on the skills requested. The Company's patented algorithm manages over 100,000 unique technology skills. Additionally, our marketplaces allow tech professionals to find their ideal next career opportunity, with relevant advice and personalized insights. Learn more at View source version on Contacts Investor ContactTodd Kehrli or Jim ByersPondelWilkinson, Inc.212-448-4181ir@
