By building as much AI data center capacity as quickly as possible, Google will see its greenhouse gas emissions increase by 48 percent in just five years.

Google’s uncontrolled appetite for more AI capacity is having a negative impact on the environment and climate, at least in the short term. Google has seen its greenhouse gas emissions increase by 48 percent in just five years and by thirteen percent compared to 2022. Net zeroGoogle’s goal is therefore at risk.

(Limited) green energy

Google uses green energy wherever possible. Greenhouse gas emissions therefore grow at a lower percentage rate than energy consumption itself. This increased by 37 percent in 2023 compared to 2022. However, the figures paint a somewhat distorted picture, as Google’s high electricity consumption means it consumes green energy that can no longer be used for other purposes.

The impact of the hunger for AI data centers goes beyond their consumption. The large infrastructure projects are also increasing emissions in the supply chain, by eight percent. In addition, some green energy projects have been stopped in 2023.

Google is not alone. Most technology companies are currently investing billions in AI. For the same reason, Microsoft has seen its emissions increase by about a third since 2022.

Global impact

The impact of these projects is immense. Google itself estimates that it consumes seven to ten percent of all the energy that goes into data centers worldwide. Data centers generally consume two to four percent of the world’s available energy.

Data center specialists argue that their large data centers are more efficient than those that companies build themselves. They also point out that the data center uses energy primarily to satisfy user demand, so they bear at least some responsibility. This is true for classic IT, but in the context of AI this argument is less strong.

Impact on objectives

Google, Microsoft, AWS, Nvidia and other major players are making a conscious decision to innovate uninhibitedly in the field of generative AI. The only factor that currently determines the development and deployment of AI is the availability of GPUs and their performance. Sustainability doesn’t really play a role, as the above numbers show. Whether and how Google and others will reconcile this reality with their sustainability goals is currently very unclear.