Is electricity the end of AI? UK power grid chief: Data center power consumption expected to increase sixfold in the next 10 years

According to the Zhītōng Finance and Economics APP, the head of the National Grid Plc in the UK stated on Tuesday local time that due to the vigorous development of artificial intelligence technology requiring a significant increase in data center computing power, it is expected that the electricity demand for UK data centers will increase by about 6 times in the next 10 years. This will significantly increase the pressure on the country's power grid system, as the grid must transport large amounts of renewable energy from distant Scottish wind farms to AI data centers around London. However, the UK power grid is already under immense pressure due to the accelerating trend of electrification in home heating, transportation, and industry.

John Pettigrew, CEO of National Grid Plc, stated at a meeting in Oxford on Tuesday, "The future development trends of foundational technologies such as artificial intelligence and quantum computing will mean larger-scale, energy-intensive computing infrastructure."

Other European countries also face similar electricity consumption prospects, as they are working to address the issue of funding the massive expenditures required for the expanding electricity generation capacity. According to the latest forecast data from the International Energy Agency (IEA), over the next three years, the electricity demand for high-consumption industries such as data centers, artificial intelligence, and cryptocurrencies may double, equivalent to the entire electricity demand of Germany, the largest economy in Europe.

In the UK, National Grid Plc is studying the addition of an ultra-high voltage power network of up to 800 kilovolts, which is approximately twice the current capacity.

CEO Pettigrew of National Grid Plc stated that the UK's new grid system will make large-scale electricity transmission across the entire UK possible, aiming to connect the sources of large-scale energy with the centers of electricity demand. This will transform from the currently highly dispersed individual connection projects - all of which require their own approvals and infrastructure - into super-large electricity capacity centers.

It is understood that this significant proposal is in addition to the previously announced approximately £112 billion (about $142 billion) UK grid investment plan to ensure the country moves towards net-zero emissions.

"This kind of project is indeed very expensive," said UK Energy Minister Graham Stuart. "The biggest obstacle my department faces in this country is building the grid."

As human society enters the AI era, electricity consumption may surge! Power resources are of paramount importance.

In 2023, ChatGPT became popular worldwide, in 2024, Sora's video large model was launched, and in the AI field, Nvidia's "shovel seller" achieved unparalleled performance for four consecutive quarters, indicating that human society will gradually enter the AI era from 2024, with power resources playing a crucial role as a core component of AI infrastructure.

Dutch National Bank data expert Alex de Vries estimates that OpenAI's chatbot ChatGPT consumes over 500,000 kilowatt-hours of electricity per day to process approximately 200 million user requests, equivalent to more than 17,000 times the daily electricity consumption of American households Google's search engine also requires a large amount of electricity. According to de Vries' predictions, if Google were to use AGI (Artificial General Intelligence) in every search, its annual electricity consumption would increase to about 29 billion kilowatt-hours, exceeding the annual electricity consumption of many countries, including Kenya, Guatemala, and Croatia.

Based on predictions from the Uptime Institute, a data center standard organization and third-party certification body, it is estimated that by 2025, the share of artificial intelligence business in global data center electricity consumption will surge from less than 2% to over 10%.

According to forecasts from the Boston Consulting Group, the share of data centers in U.S. electricity consumption is expected to double, increasing from 126 terawatt-hours in 2022 to 390 terawatt-hours in 2030. Terawatt-hours are used to describe the largest levels of electricity consumption, typically for national energy statistics, large-scale energy project planning and evaluation. Large industrial facilities, such as super steel plants, may consume less than 10 terawatt-hours of electricity in a year.

Musk predicts that in the next two years, the AI technology industry will shift from being "silicon-starved" to "electricity-starved," which could hinder the development of artificial intelligence technology. In his view, a shortage of electricity could have serious consequences, similar to how a shortage of chips will hinder the development of the human technology industry