How AI is increasing the carbon footprint

The current AI boom is putting the tech giants in a difficult position when it comes to achieving their climate protection goals. This is demonstrated by the examples of Google and Microsoft, who recently published their environmental reports. These show that developments in the field of artificial intelligence and the enormous computing power required for this are causing greenhouse gas emissions to skyrocket.

Google set itself an ambitious goal years ago. The search engine giant wants to achieve “net zero emissions” by 2030, i.e. offset all greenhouse gas emissions caused by reduction measures. According to Google, this also includes operating all of its facilities worldwide with CO2-free energy around the clock.

Figures from the company’s own environmental report published on Tuesday suggest that this could be a herculean task. There is still “a long way to go” to reach the goal of reducing direct and indirect greenhouse gas emissions by half by 2030, as the company itself admits in the report. This is also supported by the fact that Google’s carbon footprint has not become smaller in recent years – on the contrary. Since 2019, the company’s annual greenhouse gas emissions have increased by 48 percent. Last year alone, emissions amounted to 14.3 million tons – an increase of 13 percent compared to 2022.

In the report, Google also provides explanations for these rising numbers. The increased CO2 emissions are mainly due to the high energy demand in data centers on the one hand and emissions in supply chains on the other. According to a study by the International Energy Agency (IEA), data centers were responsible for one to 1.3 percent of global electricity demand in 2022. The company estimates that Google’s data centers will account for up to ten percent of this.

A look into the future

While data centers are already power guzzlers by nature, artificial intelligence significantly amplifies this effect. “While AI offers new solutions for climate protection, it also brings with it a climate impact,” the report says. Accordingly, the electricity consumption of Google’s data centers increased by 17 percent last year compared to 2022. Google assumes that “this trend will continue in the future.”

Google is gradually increasing its efforts in the field of artificial intelligence. The Gemini AI model is being integrated into more and more products, and the search engine has also been refreshed with generative AI elements. There are currently no signs that these efforts will be curtailed in the future. If you read Google’s environmental report, the opposite seems to be the case: “Our access will continue to evolve and we will have to deal with significant uncertainties – including the uncertainty around the future environmental impacts of AI, which are complex and difficult to predict.”

• Continue reading: Data centers – Why demand exceeds supply

Microsoft continues to invest

Another example shows that AI developments are good advertising for one’s own cause, but also extremely resource-intensive: Microsoft. The Windows group also recently announced in its environmental report that CO2 emissions last year were 29 percent higher than in 2020. The comparison with 2020 is explosive because Microsoft made a promise at the beginning of this year to be CO2-negative by 2030, i.e. to remove more carbon dioxide from the atmosphere than it causes.

In the meantime, the company has invested billions in the ChatGPT developer OpenAI and the development of generative artificial intelligence. And here too, it is the data centers that are responsible for a large part of the electricity consumption. There is no end in sight to this development. Just on Wednesday, Microsoft announced investments of 2.2 billion euros in data centers in the northern Spanish region of Aragon.