AI is not the answer to climate change, but the cause: it will double the planet’s energy consumption by 2030

This week, in its annual sustainability report, Google announced that its climate emissions had increased by 48% since 2019. Amazon generates as much CO2 as Switzerland (up 40% since 2019) and, for its part, Microsoft’s figures indicate an increase of 30% since 2019. And in all cases, behind These figures reflect the consumer needs of artificial intelligence.

According to data from the United Nations, energy consumption and production are by far the biggest contributors to global climate change: They account for more than 75% of global greenhouse gas emissions and almost 90% of all carbon dioxide emissions.. And it is a self-reinforcing circle. On the one hand, annual electricity consumption increases by 9.2% for every +1 °C of average global temperature, and if consumption increases, so does the temperature.

The problem with AI is that tech companies are building vast networks of data centers around the world to power this technology. And they need a lot of it. Every time a user enters a request into a chatbot or generative AI tool, the request is sent to a data center. And some of the power is consumed there to power the data center. processing information (40%) and part for cooling (40%)according to data from the International Energy Agency (IEA). And AI requires much more energy than “conventional” systems: several studies have shown that every request made to ChatGPT uses about 10 times the power of a single Google search.

So if Google switches all search queries to AI (about nine billion a year), it could increase the company’s electricity consumption by tenfold, if we follow this logic. And that’s just Google. Before the AI ​​era, estimates generally suggested that data centers accounted for about one percent of global electricity demand. The IEA report says that data centers, cryptocurrencies, and AI combined used 460 TWh of electricity worldwide in 2022, almost two percent of total global electricity demand. And the figure could double by 2026, matching Japan’s consumption.

But there is more. Data centers and their transmission networks have become the main driver of global energy consumption. This currently accounts for 3% of global consumption and emits as much CO₂ as Brazil. The growing energy needs also show no signs of slowing down, as consumption could grow from 460 terawatt-hours in 2022 to 1000 twh in 2026. And data center energy demand is expected to double by 2030. And this makes it difficult for Microsoft, Google and Amazon to target the coming years: The first two wanted to reduce their emissions to zero by 2030, while Amazon pointed to 2040 as the target date.something very complex considering that Jeff Bezos’ company is already the largest buyer of renewable energy on the planet. And even Brad Smith, president of Microsoft, admitted the difficulty of the objective: “We made the promise 4 years ago and it was a leap to the Moon, but it was made before the explosion of AI and now the Moon is five times further away than it was in 2020.”

The diagram is simple: AI will require more and more energy, energy consumption increases the temperature and the increase in temperature increases the demand for electricity. needs more energy and the wheel keeps turning, driven by AI.