Artificial intelligence could soon outgrow America’s power grid, according to Elon Musk, who warned that the increasing demand from AI data centers might overwhelm electricity supplies by the end of next year.
In a recent interview with CNBC, Musk said AI is growing so fast that it’s about to hit a shortage of computer chips, a shortage of transformers, and a looming shortage of actual power generation capacity.
He emphasized that once the chip and transformer issues are addressed, the biggest hurdle will be generating enough electricity to support this AI boom.
Musk builds a gigawatt-scale data center to meet AI power needs
Elon Musk’s xAI is building a large data center (project Colossus) outside Memphis, Tennessee, that will end in six to nine months. According to the Department of Energy, Musk explained that the facility would need one gigawatt of electricity, which matches the output of an average nuclear power plant in the United States.
xAI plans to use natural gas turbines to help generate the electricity needed to power the Colossus data center because it will support the huge computational loads required for training and running advanced AI systems.
Environmental groups claim that xAI may be violating the Clean Air Act by operating gas turbines without using pollution control technology or securing the necessary environmental permits. They urge regulators to step in because the facility could become a major source of air pollution.
The fact that xAI needs fossil fuel infrastructure to meet its power needs raises broader concerns about how the next wave of AI expansion could affect the environment, which makes Musk’s Colossus project a warning that this growth could come with significant environmental costs unless it prioritizes clean energy.
Experts warn the U.S. power grid may not keep up with AI’s explosive growth
Data centers keep growing and demanding more power every year, raising concerns from energy and tech industry leaders about whether the U.S. electric grid can support the rapid growth of artificial intelligence.
Google’s energy strategy lead, Caroline Golin, stated at a Nuclear Energy Institute conference that the company ran into a “stark reality,” which simply means they didn’t have enough electricity to power both their existing and future data centers.
Golin explained that Google started exploring nuclear power because renewable sources aren’t as reliable since they depend on weather conditions that change daily and seasonally, highlighting how even tech giants with early investments in green energy now worry that the grid won’t keep up with AI-driven growth.
Elon Musk described China’s energy expansion as “a rocket going to orbit” and said the U.S. is falling behind such countries when it comes to building new power capacity because there’s been very little growth in energy infrastructure despite surging demand.
Utility company Dominion Energy in Northern Virginia also told investors they are not seeing any slowdown in demand, which shows that more AI facilities are coming online and will require even more power.
However, the CEO of Constellation Energy, Joe Dominguez, said during an earnings call that many energy load forecasts are likely overstated because developers often present the same data center projects to multiple states, which makes it hard to get accurate estimates of real energy needs.
These mixed messages from companies building AI systems, utilities supplying the power, and energy providers debating the forecasts create a deep uncertainty about whether America’s grid can meet the rising power demands of the AI boom.
The AI industry risks running into a wall without serious investment in both clean and reliable energy sources because there simply isn’t enough electricity to keep the lights on.
Cryptopolitan Academy: Want to grow your money in 2025? Learn how to do it with DeFi in our upcoming webclass. Save Your Spot
