Greenpeace East Asia has reported a staggering increase of over 350% in global emissions from AI chip manufacturing between 2023 and 2024. The primary culprits behind this rise are the fossil fuel-powered electricity grids in Taiwan, South Korea, and Japan, which host most of the world’s AI chip production facilities. Artificial intelligence technologies have seamlessly integrated into our daily lives, with many individuals unknowingly using AI-driven products on their smart devices. This widespread adoption fuels a growing demand for semiconductor chips essential for these AI tools.

Consequently, there has been a notable surge in global electricity consumption, leading to higher carbon emissions worldwide. A recent study by Greenpeace indicates that the emissions from electricity consumed in AI chip production have skyrocketed, with a reported increase of over 350% between 2023 and 2024. In that time, electricity consumption for AI chip manufacturing has more than tripled, reaching an alarming 984 gigawatt hours (GWh). Specifically, emissions have reached approximately 453,600 metric tons.

Projections suggest that by 2030, electricity demand for AI chip production might soar up to 170 times the levels observed in 2023, surpassing the consumption of certain small nations like Ireland. The report emphasizes that major chip manufacturers, such as NVIDIA, rely heavily on suppliers like TSMC and SK Hynix, which are predominantly based in regions dependent on fossil fuels for power. If this reliance continues, the carbon footprint associated with producing these chips will only increase. The rapid growth of AI technologies raises significant environmental concerns and threatens global decarbonization efforts.

Greenpeace is urging leading tech companies, including NVIDIA, Microsoft, Meta, and Google, to support their suppliers in transitioning to renewable energy, aiming for a target of 100% renewable energy across their supply chains by 2030. While some companies like TSMC are making strides towards renewable energy, the pace remains slow. Encouragingly, Google is developing AI tools designed to enhance electricity system efficiency, and there is hope that more firms will take similar steps to mitigate their environmental impact.