AI consumes more electricity than Bitcoin.

According to research by Alex de Vries-Gao, some AI models consume as much electricity as an entire country. Photo: TheVerge .

According to a new analysis, artificial intelligence (AI) could soon surpass Bitcoin mining in terms of energy consumption. The study concludes that AI could use nearly half of the total electricity consumed by data centers globally by the end of 2025.

These estimates come from Alex de Vries-Gao, a research fellow at the Vrije Institute for Environmental Studies at Universiteit Amsterdam, where he has been tracking the electricity consumption and environmental impact of cryptocurrencies. His latest commentary on the growing electricity demands of AI was published in the journal Joule last week.

"The bigger the better"

Currently, AI is estimated to account for up to 20% of the electricity used by data centers. According to De Vries-Gao's analysis, based on supply chain forecasts for specialized AI chips (due to a lack of specific data from technology companies), this consumption is escalating rapidly, despite improvements in efficiency.

De Vries-Gao once thought research into energy-intensive technologies might end with Ethereum The Merge. However, the emergence of ChatGPT has directed his research focus to a new direction. The analyst sees striking similarities between the development of AI and the energy demands of cryptocurrency markets.

"When ChatGPT came out, I thought, 'Oh my god, again.' This is another technology that consumes a lot of energy, especially in highly competitive markets," he shared with The Verge .

The AI industry will increasingly demand more energy due to competition. Photo: SciTechDaily.

A key commonality is the prevailing "bigger is better" mindset in both industries. "We see tech companies constantly scaling up their models to create the best possible applications. However, this also increases the demand for resources," De Vries-Gao explains.

The pursuit of this trend has fueled a boom in new AI data centers, particularly in the US. This has led to plans to build more gas-fired power plants and nuclear reactors to meet the growing electricity demand.

Sudden surges in electricity demand can put significant pressure on the grid and hinder the transition to cleaner energy sources, similar to the challenges posed by cryptocurrency mining. Another similarity is the difficulty in accurately assessing the electricity consumption and environmental impact of these technologies. While big-tech companies report carbon emissions, they rarely provide specific data on their own AI.

To solve this problem, De Vries-Gao used a "triangulation" technique. He used publicly available information about the devices, estimates from analysts, and revenue reports from companies to forecast the number of hardware units to be produced and the amount of energy they might consume.

He also noted that Taiwan Semiconductor Manufacturing Company (TSMC), a major AI chip manufacturer, has more than doubled its AI chip production capacity from 2023 to 2024.

Forecasts and uncertainties ahead

De Vries-Gao estimates that by 2024, AI was consuming the equivalent of the entire Netherlands' electricity. By the end of 2025, this figure could rise to match that of the United Kingdom, with AI's electricity demand reaching 23 gigawatts (GW).

A separate report from the consulting firm ICF also forecasts that electricity demand in the US will increase by 25% by 2030. This increase is primarily driven by AI, data centers, and Bitcoin mining.

Despite these predictions, providing an accurate figure for AI's energy consumption remains complex. Environmental impact varies significantly depending on many factors such as the type of processing required, the size of the AI model, and the power supply to the local grid.

Ethereum's electricity consumption has decreased by 99.988% after switching to a more energy-efficient transaction validation method compared to Bitcoin. Photo: SCMP.

For example, using AI tools processed by data centers in West Virginia can generate nearly twice the carbon emissions compared to California. This is due to differences in renewable energy use between the two states.

De Vries-Gao believes that tech companies need more transparency. "Having to go through so many complicated steps to arrive at an estimate is truly absurd. It shouldn't be ridiculously difficult, but sadly it is," he shared.

Looking to the future, whether energy efficiency will increase remains an open question. While some AI models, such as DeepSeek's, claim significantly lower power consumption than others, the question is whether companies will prioritize efficiency over the "bigger is better" trend.

The risk of the Jevons paradox—where increased efficiency leads to higher overall consumption due to increased usage—is also present. Without better metrics and transparency, managing the energy consumed by AI will be a major challenge.

Source: https://znews.vn/ai-co-the-tieu-thu-dien-nhieu-hon-bitcoin-vao-cuoi-nam-2025-post1556958.html