Google, a leading tech giant, has been aggressively integrating artificial intelligence (AI) into its core products, such as Google Search and Google Assistant. However, this AI integration has led to a significant increase in the company’s greenhouse gas emissions. The reason for this increase is the energy consumption of the data centers required to run AI systems, which produce a large amount of heat and require substantial amounts of electricity.
According to Google’s annual environment report, the company’s greenhouse gas emissions have risen by 48% since 2019, primarily due to increased data center energy consumption and supply chain emissions. Google has acknowledged that reaching its net-zero emissions goal by 2030 will be challenging due to the complex and unpredictable environmental impact of AI.
The International Energy Agency estimates that a Google search query requires 0.3 watt-hours of electricity on average, while a ChatGPT request consumes about 2.9 watt-hours, illustrating the power-hungry nature of AI models. Google expects its total greenhouse gas emissions to continue to rise before falling as it invests in clean energy sources like wind and geothermal energy to power its data centers.
In addition to the energy consumption, the large amounts of water used as coolant in data centers also pose a sustainability challenge. Google aims to replenish 120% of the freshwater it consumes in its offices and data centers by 2030, but it only managed to replenish 18% of that water in 2023, although the amount was up sharply from 6% the year prior.
Despite the environmental challenges associated with AI, Google recognizes the importance of scaling AI and using it to accelerate climate action. The company has already used AI to suggest more fuel-efficient routes to drivers using Google Maps and to train an AI model to predict the availability of wind power, helping to increase the value of renewable energy for wind farmers. Google is among the companies experimenting with ways to use AI to fight climate change.
