A study by PhD candidate Alex de Vries-Gao revealed that AI could soon surpass Bitcoin mining in energy consumption. The new analysis claimed that AI could use nearly half of all the power consumed by data centers globally by the end of 2025.
De Vries-Gao’s study concluded that AI could be drawing as much electricity as some mid-sized European nations by the end of this year, and the numbers are only rising. AI already accounts for up to a fifth of the power used by data centers, but it is still tricky to pin down the exact figures without big tech companies sharing data on how much energy their AI models consume.
In 2023 and 2024, the adoption of generative AI applications drew media attention to the increasing power demand for these applications. Tech companies faced a “power capacity crisis” in their efforts to expand data center capacity as power demand rose to support these applications.
De Vries-Gao makes projections based on the ‘triangulation’ technique
AI has a power problem. While tech giants pour billions into development, they're hitting a wall that has nothing to do with chips, algorithms, or talent. The problem is fundamental: the electrical grid can't handle what's coming. The latest Prinsights linked below details why. pic.twitter.com/0L4JXJgebV
— Nomi Prins (@nomiprins) May 28, 2025
De Vries-Gao made projections based on the supply chain for specialized computer chips used for AI as he and other researchers tried to understand AI’s energy consumption. The researchers discovered that AI’s energy appetite is growing fast enough to warrant more scrutiny despite efficiency gains.
To figure this out, De Vries-Gao used what he called a “triangulation” technique.
The researcher turned to publicly available device details, analyst estimates, and companies’ earnings calls to estimate hardware production for AI and how much energy that hardware is likely to use.
De Vries-Gao calculated how specialized AI equipment could be produced and compared that to information about how much electricity these devices consumed. He found that last year, they likely burned through as much electricity as his home country of the Netherlands.
“That’s how we know that both Google’s and Microsoft’s carbon footprints have grown in recent years as they focus on AI. But companies usually don’t break down the data to show what’s attributable to AI specifically.”
–Alex de Vries-Gao, the founder of Digiconomist
Last week, a separate report from consulting firm ICF forecasted a 25% rise in energy demand in the U.S. by the end of the decade, largely thanks to AI, traditional data centers, and Bitcoin mining.
Another article published in MIT Technology Review last week, with support from the Tarbell Center for AI Journalism, claimed that it is still hard to make blanket predictions about AI’s energy consumption and the resulting environmental impact.
Meta plans to power a giant AI data center with gas
Meta announced in mid-May that it is building a new AI data center so large in Louisiana that the local utility company planned to set up three new gas-fired power plants to provide it with enough energy. However, advocates and lawmakers pressed Meta for answers about how it would clean up pollution from the data center’s energy consumption.
Sen. Sheldon Whitehouse (D-RI) wrote a letter to Meta CEO Mark Zuckerberg on May 14 demanding answers about how much energy the data center would use. He also sought answers about the greenhouse gas emissions that would be generated. The letter pointed out that the gas-powered new data center flew in the face of Meta’s climate commitments.
Sen. Whitehouse also stated that Meta’s backslide from its own climate pledges in 2020 risks triggering broader economic harm at a time when corporate responsibility is urgently needed.
Logan Burke, the Executive Director of the Alliance for Affordable Energy, claimed it is hard to figure out whether a facility like Meta’s is good or bad for communities without understanding the possible impact on electrical systems, bills, and water.
However, Meta maintained that it will continue matching its electricity use with support for renewable energy, including a commitment to help fund 1,500 megawatts of new solar and battery resources in Louisiana.
Join a premium crypto trading community free for 30 days - normally $100/mo.
