Artificial intelligence's growing emissions footprint is becoming glaringly evident, with Google reporting a 13 percent rise in emissions due to increased data centre electricity use, driven by expanding AI deployments
The environmental toll of artificial intelligence (AI) is escalating, revealing a massive emissions footprint that is becoming increasingly apparent. Google’s latest annual environment report shows a concerning 13 percent rise in its emissions for 2023, primarily driven by a surge in electricity usage at its data centres, which saw a 17 percent increase.
This uptick in energy consumption is closely linked to the expanding deployment of AI tools. AI’s energy demands are starkly higher than traditional tech uses. A single AI query, such as those processed by OpenAI’sChatGPT, can consume up to 33 times more energy than a standard Google search. This inefficiency is compounded by the extensive data processing required for AI operations, which not only increases electricity use but also generates more heat—the result is higher energy consumption for cooling data centres.
The global impact of this trend is significant. Data centres currently account for 1 percent to 1.3 percent of global electricity demand, with projections indicating this could double by 2026, according to the International Energy Agency (IEA). This contrasts sharply with electric vehicles, which use about 0.5 percent of global electricity. In some regions, the share of national electricity demand consumed by data centres has already exceeded 10 percent.
Saurabh Rai, CEO of Arahas Technologies, warns that India will soon confront the severe environmental costs associated with AI and data centres.