Data centers would rank 17th in global energy use and demand is skyrocketing

Keywords: Energy, Learn, Newsroom

๐—œ๐—ณ ๐—ฑ๐—ฎ๐˜๐—ฎ ๐—ฐ๐—ฒ๐—ป๐˜๐—ฒ๐—ฟ๐˜€ ๐˜„๐—ฒ๐—ฟ๐—ฒ ๐—ฎ ๐—ฐ๐—ผ๐˜‚๐—ป๐˜๐—ฟ๐˜†, ๐˜๐—ต๐—ฒ๐˜†โ€™๐—ฑ ๐—ฟ๐—ฎ๐—ป๐—ธ 17๐˜๐—ต ๐—ถ๐—ป ๐˜„๐—ผ๐—ฟ๐—น๐—ฑ๐˜„๐—ถ๐—ฑ๐—ฒ ๐—ฒ๐—ป๐—ฒ๐—ฟ๐—ด๐˜† ๐—ฐ๐—ผ๐—ป๐˜€๐˜‚๐—บ๐—ฝ๐˜๐—ถ๐—ผ๐—ป. ๐—”๐—ป๐—ฑ ๐˜๐—ต๐—ฒ๐—ถ๐—ฟ ๐—ฒ๐—ป๐—ฒ๐—ฟ๐—ด๐˜† ๐—ป๐—ฒ๐—ฒ๐—ฑ๐˜€ ๐˜„๐—ถ๐—น๐—น ๐—ธ๐—ฒ๐—ฒ๐—ฝ ๐—ฟ๐—ถ๐˜€๐—ถ๐—ป๐—ด ๐—ฟ๐—ฎ๐—ฝ๐—ถ๐—ฑ๐—น๐˜†.

Andreas Horn

The rapid growth of data centers is seriously undermining global CO2 reduction efforts. Here are some key facts highlighting their environmental impact:

  • Current Consumption: Data centers currently consume 1-2% of global power, with projections suggesting this could rise to 3-4% by the end of the decade.
  • Carbon Emissions: These emissions are expected to more than double between 2022 and 2030.
  • Rapid Growth: While workloads nearly tripled from 2015 to 2019, power demand has remained steady at around 200 terawatt-hours per year.
  • AI Impact: A single ChatGPT query uses 2.9 watt-hours, compared to 0.3 watt-hours for a Google search. By 2030, AI could add 200 terawatt-hours per year to data center power consumption.
  • Carbon Footprint: Data centers now emit more carbon than all global airline travel combined, surpassing most major nations.
  • Social Cost: The increase in data center carbon emissions could impose a social cost of $125-140 billion in present value.

Europe’s Challenge:

  • Aging Power Grid: Europe, with the oldest power grid globally, faces significant challenges in keeping new data centers powered. A recent Goldman Sachs study indicates that over โ‚ฌ1 trillion ($1.1 trillion) will be required for grid upgrades and renewable energy investments.
  • Rising Power Demand: European power demand is expected to grow by 40-50% between 2023 and 2033.
  • Future Needs: By 2030, European data centers, which currently house 15% of global capacity, will require power equivalent to the combined needs of Portugal, Greece, and the Netherlands.

As the push towards advanced AI technologies accelerates, energy consumption is set to soar, with training new models becoming increasingly energy-intensive.